Beyond Standardization: Advancing Nutrient Profiling Models for Traditional and Modern Food Varieties in Biomedical Research

Abstract

This article provides a comprehensive analysis of nutrient profiling models (NPMs) and their critical application in evaluating traditional and modern food varieties for drug development and biomedical research. It explores the foundational science behind NPMs, from static, population-based systems to dynamic, AI-enhanced platforms. The content details methodological approaches for adapting NPMs to diverse food systems, addresses key challenges in cultural adaptation and data integration, and presents frameworks for validating models against health outcomes. Aimed at researchers and scientists, this review synthesizes current evidence to guide the selection, optimization, and application of robust nutrient profiling tools in clinical and product development settings, ultimately bridging the gap between nutritional science and precision medicine.

From Static Scores to Dynamic Systems: The Evolution of Nutrient Profiling Science

Nutrient profiling (NP) is defined as "the science of classifying or ranking foods according to their nutritional composition for reasons related to preventing disease and promoting health" [1]. This scientific discipline provides objective, transparent, and reproducible methods to evaluate the nutritional quality of individual foods and beverages based on their nutrient composition [2]. Nutrient profiling models (NPMs) utilize specific algorithms that consider the amounts of multiple nutrients and food components to characterize a product's "healthfulness" through numerical scores or qualitative classifications [2].

These models serve as crucial tools in nutritional research, particularly when comparing traditional versus modern crop varieties. Research indicates that modern plant varieties often "perform worse in terms of organoleptic features, especially of volatile components content" compared to old traditional cultivars and their wild relatives [3]. The application of NP models enables researchers to quantitatively assess these nutritional differences, providing evidence-based insights into how agricultural practices and crop selection influence food quality and human health.

Core Principles of Nutrient Profiling

Foundational Concepts

The fundamental principle underlying nutrient profiling is the concept of nutrient density – identifying foods that supply relatively more beneficial nutrients than calories [4]. Effective NP models are characterized by several core principles: they must be transparent, based on publicly accessible nutrient composition data, and validated against independent measures of a healthy diet [4]. The development of robust NP models follows rigorous scientific standards, including repeated testing against established dietary quality indices such as the Healthy Eating Index (HEI) [4].

Most NP models incorporate two key dimensions: nutrients to encourage (e.g., protein, fiber, vitamins, minerals) and nutrients to limit (e.g., saturated fat, added sugar, sodium) [4] [2]. This balanced approach ensures comprehensive nutritional assessment rather than focusing exclusively on negative or positive aspects. The selection of specific nutrients and the weighting assigned to each component varies depending on the model's intended application and the nutritional priorities of the target population.

Model Structures and Calculation Methods

NP models employ different structural approaches and calculation methods, each with distinct advantages:

• Scoring Systems: Models calculate nutritional quality based on nutrient content per 100g, 100kcal, or per serving size [4]. Research indicates that models based on 100kcal and serving sizes generally demonstrate better performance than those based on 100g alone [4].

• Algorithm Types: Formulas based on sums and means typically outperform those based on ratios [4]. For instance, the Nutrient-Rich Foods (NRF) family of models uses a combined approach that sums beneficial nutrients and subtracts limiting nutrients [4].

• Classification Systems: Some models utilize categorical classifications (e.g., "healthy"/"unhealthy") rather than continuous scores, which can simplify implementation for policy applications [2].

Table 1: Common Nutrient Profiling Model Types and Characteristics

Model Type	Key Features	Common Applications	Examples
Threshold Models	Sets specific limits for nutrients	Front-of-pack labeling, marketing restrictions	PAHO Model, Choices International
Scoring Models	Calculates continuous scores based on multiple nutrients	Food ranking, dietary guidance	NRF Index, Food Compass
Hybrid Models	Combines scoring with categorical classification	Product reformulation, consumer education	Nutri-Score, Health Star Rating

Analytical Techniques for Nutritional Composition Analysis

Accurate nutrient profiling depends on precise analytical techniques to determine food composition. Several advanced methodologies enable comprehensive nutritional assessment.

Chromatographic Methods

Gas Chromatography (GC) is a widely used technique for separating and analyzing compounds that can be vaporized without decomposition. GC operates by passing a vaporized sample through a column with an inert gas, separating components based on their interactions with the stationary phase [1]. The retention index is calculated using the equation:

[ I = 100z + 100 \frac{\log(t{R(x)}^1 - t{R(z)}^1)}{\log(t{R(z+1)} + t{R(z)}^1)} ]

where ( t'_{R} ) represents adjusted retention time, and ( z ) denotes the number of carbon atoms in reference hydrocarbons [1]. GC applications in food variety analysis include assessing sterols, oils, low-chain fatty acids, aroma components, and contaminants such as pesticides [1].

Liquid Chromatography techniques, including High-Performance Liquid Chromatography (HPLC), are employed for analyzing water-soluble vitamins, amino acids, and other polar compounds. These methods are particularly valuable for quantifying specific bioactive compounds in traditional crop varieties that may be lost during modern breeding programs [1].

Emerging Analytical Technologies

Advanced analytical approaches are enhancing the depth and accuracy of nutrient profiling:

• Metabolomics provides comprehensive analysis of small molecule metabolites, enabling detailed characterization of the biochemical composition of different crop varieties [1]. This approach has revealed that traditional rice varieties contain "over 149 metabolites grouped into 34 chemical classes," demonstrating their complex biochemical profiles [3].

• Proteomic and Molecular Assays facilitate the identification and quantification of proteins and other macromolecules, contributing to understanding the functional properties of food components beyond basic nutrition [1].

• Nanotechnology Applications offer enhanced sensitivity and specificity for detecting micronutrients and bioactive compounds at minimal concentrations, enabling more precise nutritional comparisons between food varieties [1].

Table 2: Analytical Techniques for Nutrient Profiling of Food Varieties

Technique	Applications	Sensitivity	Throughput
Gas Chromatography	Fatty acids, sterols, aroma compounds	High	Medium
Liquid Chromatography	Water-soluble vitamins, amino acids	High	Medium-High
Mass Spectrometry	Metabolite profiling, micronutrient analysis	Very High	Medium
Spectrophotometry	Macronutrient analysis, antioxidant capacity	Medium	High

Experimental Protocols for Food Variety Analysis

Protocol 1: Comprehensive Metabolite Profiling of Crop Varieties

Purpose: To systematically identify and quantify metabolic differences between traditional and modern crop varieties.

Materials and Reagents:

• Freeze-dried tissue samples from multiple varieties
• Methanol, acetonitrile, and water (HPLC grade)
• Derivatization reagents (for GC-MS analysis)
• Internal standards (e.g., stable isotope-labeled compounds)
• Solid-phase extraction cartridges

Procedure:

• Sample Preparation: Homogenize 100mg of freeze-dried plant material in liquid nitrogen. Extract metabolites using 1mL of methanol:water (4:1, v/v) with sonication for 30 minutes at 4°C [1].
• Metabolite Extraction: Centrifuge at 14,000 × g for 15 minutes. Collect supernatant and evaporate under nitrogen stream. Reconstitute in appropriate solvent for instrumental analysis.
• Instrumental Analysis:
  • For GC-MS: Derivatize samples using N-methyl-N-(trimethylsilyl)trifluoroacetamide (MSTFA). Analyze using DB-5MS column with temperature programming from 60°C to 330°C [1].
  • For LC-MS: Use reversed-phase C18 column with gradient elution (0.1% formic acid in water and acetonitrile).
• Data Processing: Use specialized software (e.g., XCMS, MS-DIAL) for peak detection, alignment, and annotation against spectral databases (e.g., NIST, HMDB).

Validation: Include quality control samples (pooled quality control) throughout the analysis sequence to monitor technical variability. Validate compound identifications using authentic standards when available.

Protocol 2: Nutrient Density Scoring Using Established NP Models

Purpose: To apply standardized nutrient profiling models to compare the nutritional quality of different food varieties.

Materials:

• Comprehensive nutrient composition data for each variety
• Standardized NP model algorithms (e.g., NRF9.3, Food Compass 2.0)
• Nutrient database software (e.g., USDA FoodData Central, FooDB)

Procedure:

• Data Collection: Compile analytical data for key nutrients from at least three independent samples per variety. Essential nutrients include protein, fiber, vitamins A, C, and E, calcium, iron, magnesium, potassium, saturated fat, added sugar, and sodium [4].
• Calculation Method:
  • For the NRF9.3 model: Calculate the sum of percentage daily values (%DV) for 9 beneficial nutrients per 100kcal, then subtract the sum of %DV for 3 limiting nutrients [4].
  • For Food Compass 2.0: Score foods across 9 domains including nutrient ratios, food ingredients, and processing characteristics, calculated per 100kcal [5].
• Statistical Analysis: Perform ANOVA or multivariate analysis to determine significant differences between variety groups. Calculate correlation coefficients between NP scores and specific bioactive compounds.

Interpretation: Higher scores indicate superior nutrient density. Research applications should consider both statistical significance and practical nutritional relevance of observed differences.

Research Reagent Solutions for Nutrient Profiling

Table 3: Essential Research Reagents for Nutrient Profiling Studies

Reagent/Category	Function	Application Examples
Chromatography Standards	Quantification and method validation	Fatty acid methyl esters, amino acid mixtures, vitamin standards
Extraction Solvents	Metabolite liberation and stabilization	HPLC-grade methanol, acetonitrile, chloroform-methanol mixtures
Derivatization Reagents	Volatilization for GC analysis	MSTFA, TMS, BSTFA for GC-MS analysis
Enzyme Assay Kits	Specific nutrient quantification	Phytate, antioxidant capacity, dietary fiber assays
Internal Standards	Correction for analytical variability	Stable isotope-labeled amino acids, lipids, carbohydrates
Mobile Phase Additives	Chromatographic separation enhancement	Formic acid, ammonium acetate, ion-pairing reagents

Application to Traditional vs. Modern Crop Varieties

Nutritional Differences Between Crop Types

Research consistently demonstrates significant nutritional differences between traditional and modern crop varieties. Traditional cultivars and their wild relatives often possess "higher nutritional properties" and valuable genetic traits that have been diminished in modern varieties [3]. For example, studies of traditional rice varieties reveal "high biochemical complexity" in their metabolome, supporting "future exploitation of genetic materials in nutritional targeted breeding" [3].

The nutritional divergence between traditional and modern varieties extends beyond macronutrients to include bioactive compounds. Wild edible plants frequently contain higher concentrations of volatile compounds, including "37 different compounds, mainly monoterpenoids and benzoids," contributing to both aroma and potential health benefits [3]. This phytochemical richness represents an important dimension of nutrient density that conventional NP models may overlook.

Adaptation to Climate Change and Sustainability

Traditional crop varieties offer advantages beyond immediate nutritional value, including adaptability to environmental challenges. The "heterogeneity of traditional plant cultivars" provides substantial importance for food security, as "diverse plant populations have higher resilience and adapting ability to unfavorable biotic and abiotic conditions" [3]. This resilience is increasingly crucial in the context of climate change, where the nutritional quality of crops is threatened by changing growing conditions.

The integration of traditional varieties into modern agricultural systems presents certain challenges, including "limitations in terms of yield quantity and of storage life" [3]. However, their "added value is complex, consisting of traditional aspects, regional culinary importance, environmental adaptability, and nutritional richness," justifying conservation and utilization efforts [3].

Validation and Implementation Framework

Validation Against Health Outcomes

The criterion validity of NP models – assessing whether foods rated as healthier by the model associate with better health outcomes – is essential for establishing their utility in research and policy [6]. Validation studies have demonstrated that higher nutritional quality, as defined by established NP models, correlates with significantly lower risks of cardiovascular disease, cancer, and all-cause mortality [6].

For instance, the Nutri-Score model shows substantial criterion validation evidence, with highest compared to lowest diet quality associated with hazard ratios of 0.74 for cardiovascular disease and 0.75 for cancer [6]. Similarly, the updated Food Compass 2.0 demonstrates strong predictive validity, with each standard deviation increase in score associated with more favorable BMI, blood pressure, cholesterol levels, and biomarkers of glucose metabolism [5].

Implementation in Research and Policy

Effective implementation of NP models in food variety research requires careful model selection based on study objectives. Researchers must consider whether to adopt existing models or develop customized approaches specific to traditional crop comparisons. Key considerations include:

• Alignment with Research Questions: Models emphasizing micronutrients may be preferable for studies in populations with deficiency risks, while models focusing on limiting nutrients may better address overnutrition concerns [7] [8].

• Context Appropriateness: NP models initially developed for Western populations with overnutrition issues may require adaptation for studying traditional crops in low- and middle-income countries, where "inadequate intakes of vitamin A, B vitamins, folate, calcium, iron, iodine, and zinc" persist [7].

• Technical Feasibility: The availability of comprehensive nutrient composition data for traditional foods often presents a practical limitation, necessitating strategic prioritization of analytical efforts.

Diagram 1: Nutrient Profiling Workflow for Food Variety Analysis. This flowchart illustrates the systematic process for comparing nutritional quality between traditional and modern crop varieties, from sample preparation to research application.

Nutrient profiling provides an essential methodological framework for quantitatively assessing and comparing the nutritional quality of traditional versus modern crop varieties. The core principles of NP – including transparency, scientific validation, and balanced consideration of both beneficial and limiting nutrients – establish a robust foundation for objective nutritional assessment. The experimental protocols and analytical techniques detailed in this document enable researchers to generate comparable, high-quality data on the nutritional composition of diverse food varieties.

The application of validated NP models to traditional crop research has demonstrated the superior nutritional profiles of many heritage varieties, highlighting their potential to address both micronutrient deficiencies and diet-related chronic diseases. As agricultural systems face increasing challenges from climate change and environmental degradation, the nutritional resilience offered by traditional crop diversity becomes increasingly valuable. Future research should continue to refine NP models to better capture the full spectrum of bioactive compounds in traditional foods while developing more context-appropriate approaches for different agricultural and nutritional environments.

The science of nutrient profiling (NP), defined as the classification of foods based on their nutritional composition for preventing disease and promoting health, has undergone a significant evolution [9]. This field has shifted from a one-size-fits-all approach, which applied uniform criteria across diverse populations and food products, toward the development of personalized and context-specific models. Initially, NP models were largely developed in and for high-income countries, focusing predominantly on preventing obesity by penalizing foods high in calories, fat, sugar, and salt [7]. While this approach addressed issues of nutrient excess, it was often misaligned with the public health needs of low- and middle-income countries (LMICs), where hunger, undernutrition, and micronutrient deficiencies remain pressing concerns [7]. The historical trajectory of NP is characterized by this reckoning with geographic, economic, and dietary diversity, driving the need for more tailored profiling systems. This shift is further underscored by the transition from simple graphic tools, like food pyramids, to more nuanced plate models and, ultimately, to sophisticated scoring systems capable of integrating a wider array of health-relevant food properties [10] [5]. This article details this paradigm shift, providing application notes and experimental protocols for researchers engaged in profiling traditional versus modern food varieties.

Application Notes: The Evolution of Key Models and Concepts

From Global Hegemony to Regional Adaptation

The initial one-size-fits-all approach is exemplified by the global propagation of early models like the UK's Ofcom model, which influenced many subsequent systems [11]. A pivotal moment was the introduction of the US MyPlate model in 2011, which initiated a global trend away from food pyramids toward more intuitive plate models [10]. However, countries quickly recognized the need for adaptation. As a result, while the Polish Healthy Eating Plate and the UK's Eatwell Guide are structurally similar to MyPlate, they incorporate critical regional differences, such as recommendations on physical activity, fat quality, and beverage consumption, which the original US model lacked [10]. This demonstrates a key evolutionary step: the structural format may be shared, but the specific nutritional advice is tailored to local eating habits, available foods, and public health priorities [10].

Incorporating Scientific Advancements into Model Design

Modern NP models have expanded their scope beyond a narrow set of "nutrients to limit." The original Food Compass was a landmark development, designed to capture nine holistic domains, including nutrient ratios, food ingredients, and processing characteristics [5]. Its recent update to Food Compass 2.0 illustrates the dynamic nature of modern NP, incorporating new scientific evidence. Key improvements include:

• Accounting for Food Processing: Rather than only penalizing ultra-processed foods, Food Compass 2.0 now awards positive points for non-ultraprocessed foods, better reflecting evidence on the health benefits of minimally processed foods [5].
• Refining Dairy Fat Assessment: The model was updated to account for research suggesting relatively neutral health effects of dairy fat [5].
• Incorporating Additive Data: The scoring of artificial additives, an existing attribute, was finally enabled with newly available data, leading to lower scores for highly processed foods with multiple artificial additives [5].

This continuous refinement cycle ensures that NP models remain at the forefront of nutritional science.

Addressing the LMIC Context

A major critique of early NP models was their inadequacy for LMICs. Energy-driven models designed to combat overnutrition could inadvertently penalish energy- and nutrient-dense foods vital for addressing undernutrition [7]. The manual for developing global NP models emphasizes that models intended for LMICs must prioritize addressing inadequate intakes of vitamin A, B vitamins, folate, calcium, iron, iodine, zinc, and high-quality protein [7]. This represents a fundamental shift from a model focused solely on "nutrients to limit" to one that also emphasizes "qualifying nutrients" and "components to encourage," aligning the NP model with the most relevant public health problems.

Table 1: Comparative Analysis of Historical vs. Modern Nutrient Profiling Approaches

Feature	One-Size-Fits-All Approach (Historical)	Personalized/Contextual Approach (Modern)
Primary Public Health Focus	Preventing obesity and overnutrition (Nutrients to limit)	Dual burden of disease; also addresses undernutrition and micronutrient deficiencies (Nutrients to encourage) [7]
Geographic Applicability	Developed for High-Income Countries, applied globally	Adapted to regional needs, food supplies, and dietary traditions [10] [7]
Key Model Components	Calories, total fat, saturated fat, sugar, salt	Expands to include fiber, protein, specific vitamins/minerals, food processing, and beneficial ingredients [5]
Graphical Representation	Food Pyramids, simple plates (e.g., MyPlate)	Detailed plates (e.g., Eatwell Guide, Polish Plate), advanced scoring systems (e.g., Food Compass) [10] [5]
Model Validation	Often implemented without comprehensive validation [11]	Emphasis on validation against health outcomes and other models (e.g., Food Compass 2.0) [5]

Table 2: Key Nutrient Profiling Models and Their Characteristics

Model Name (Region)	Reference Base	Nutrients/Components Considered	Key Characteristics & Purpose
Ofcom (UK) [11]	100g	Energy, Sat. fat, Na, total sugar; Protein, fiber, FVNL*	Reference model for marketing restrictions; uses a scoring system.
Nutri-Score (France) [11] [12]	100g	Energy, Sat. fat, Sugars, Na; Protein, fiber, FVNL, nuts	A/F to E color-grade; used for front-of-pack labeling in several EU countries.
FSANZ (Australia/NZ) [11]	100g or mL	Energy, Sat. fat, Na, total sugar; Protein, fiber, FVNL	High agreement with Ofcom; used for health claims.
HCST (Canada) [11]	Serving	Sat. fat, Sugars, Na; (No beneficial nutrients)	Tiered system; fair agreement with Ofcom; focuses only on nutrients to limit.
Food Compass (International) [5]	100 kcal	9 domains including nutrient ratios, vitamins, minerals, ingredients, additives, processing	Comprehensive scoring (1-100); validated against health outcomes; suitable for mixed meals and diets.

*FVNL: Fruits, vegetables, nuts, and legumes

Experimental Protocols

Protocol 1: Validating a Nutrient Profiling Model Against Health Outcomes

This protocol outlines the methodology for establishing construct validity, as demonstrated in the validation of the Food Compass 2.0 model [5].

1. Objective: To assess the association between a Nutrient Profiling Model score and objective health biomarkers and disease prevalence in a population.

2. Reagents & Materials:

• National Health and Nutrition Survey Dataset: A large, representative dataset with detailed dietary intake data (e.g., 24-hour recalls) and linked health measurements (e.g., NHANES). [5]
• Statistical Software: R, SAS, or Stata for complex statistical modeling.
• Nutrient Composition Database: A comprehensive database (e.g., USDA Branded Food Products Database) to calculate food component scores. [7]

3. Procedure: Step 1: Calculate Individual Food Scores. Apply the NP model algorithm (e.g., Food Compass) to each food and beverage in the dietary database to generate a score for each item (FCS). [5] Step 2: Compute Overall Dietary Score. For each participant in the survey, calculate an energy-weighted average of all FCSs for the foods and beverages they consumed. This generates an individual score (i.FCS). [5] Step 3: Correlate with Dietary Pattern Metric. Calculate a correlation coefficient (e.g., Pearson's r) between the i.FCS and a validated measure of a healthy dietary pattern (e.g., Healthy Eating Index-2015) to establish convergent validity. [5] Step 4: Conduct Multivariable Regression Analysis. Model the association between the i.FCS (independent variable) and a series of health outcomes (dependent variables), including:

• Continuous biomarkers: Body Mass Index (BMI), systolic and diastolic blood pressure, LDL-cholesterol, HDL-cholesterol, HbA1c. [5]
• Binary disease outcomes: Metabolic syndrome, cardiovascular disease, cancer (ascertained via medical records or self-report with validation). [5] Step 5: Adjust for Covariates. All models must be adjusted for potential confounders such as age, sex, total energy intake, smoking status, physical activity level, and socioeconomic status. [5] Step 6: Evaluate All-Cause Mortality. Use Cox proportional hazards models to assess the relationship between i.FCS quintiles and all-cause mortality risk over a follow-up period. [5]

Protocol 2: Comparing Nutritional Profiles of Traditional vs. Modern Food Products

This protocol is designed for empirical comparison of traditional and modern food varieties, as applied in studies of traditional vs. western dietary patterns and plant-based meats. [13] [12]

1. Objective: To quantitatively compare the nutritional composition of traditional and modern food products available on the market.

2. Reagents & Materials:

• Product Sample Set: A representative selection of traditional and modern alternative products from major supermarket chains. For plant-based vs. meat comparison, include multiple product categories (e.g., burgers, sausages). [12]
• Data Collection Tool: Standardized form for recording brand, product name, ingredient list, and nutritional data per 100g. [12]
• Laboratory Equipment: If conducting independent analysis, equipment for proximate analysis (fat, protein, moisture, ash), HPLC for sugars, GC for fatty acids, and ICP-MS for minerals.

3. Procedure: Step 1: Sample Identification and Collection. Systematically identify and collect traditional and modern alternative products from predefined retail outlets. Ensure products are matched by type (e.g., beef burger vs. plant-based burger). [12] Step 2: Data Extraction. For each product, extract the following nutritional parameters from the packaging: energy (kcal), total fat, saturated fat, unsaturated fat, carbohydrates, total sugars, protein, dietary fiber, and salt (all in g/100g). [12] Step 3: Data Entry and Management. Create a structured database (e.g., in Excel or SPSS) and enter all collected data. Perform double-entry verification to ensure accuracy. Step 4: Statistical Analysis. Conduct statistical analysis using t-tests or ANOVA to compare means of nutritional parameters between the traditional and modern product groups. For dietary pattern analysis, use principal component analysis (PCA) or cluster analysis to derive patterns from food group intake data. [13] [12] Step 5: Apply NP Models. Calculate the nutrient profile score for each product using one or more NP models (e.g., Nutri-Score, Food Compass). Statistically compare the scores between the traditional and modern product groups.

Mandatory Visualizations

Diagram 1: Conceptual Evolution of Nutrient Profiling

Diagram Title: NP Model Evolution

Diagram 2: Workflow for Model Validation Against Health Outcomes

Diagram Title: Health Validation Workflow

The Scientist's Toolkit: Research Reagent Solutions

Table 3: Essential Reagents and Resources for Nutrient Profiling Research

Research Reagent / Resource	Function / Application	Examples / Specifications
National Food Composition Database	Provides standardized nutrient data for calculating model scores and analyzing dietary intake.	USDA Branded Food Products Database (BFPDB); German Bundeslebensmittelschlüssel (BLS); FAO/INFOODS tables. [13] [7]
Validated Dietary Assessment Tool	Collects individual-level food consumption data for evaluating diets and validating models.	24-hour dietary recalls (e.g., using Globodiet software); Food Frequency Questionnaires (FFQs). [13]
Nutrient Profiling Algorithm Software	Executes the mathematical calculations of complex NP models to generate scores for foods and diets.	Custom scripts in R or Python to implement models like Food Compass, Nutri-Score, or Ofcom. [11] [5]
Population Health & Consumption Dataset	Provides linked dietary and health data for validating NP models against real-world outcomes.	Datasets like NHANES (US), with biomarkers, clinical data, and mortality follow-up. [5]
Statistical Analysis Package	Performs multivariate analyses, factor analysis, and regression modeling for pattern derivation and hypothesis testing.	SAS, Stata, R, SPSS, with capabilities for principal component analysis (PCA) and cluster analysis. [13] [5]

Nutrient profiling (NP) is defined as the science of classifying foods according to their nutritional composition for purposes of promoting health and preventing disease [11]. As regulatory bodies and researchers seek to improve dietary quality, two dominant philosophical approaches have emerged: "Nutrients to Limit" (NTL) models, which focus primarily on reducing consumption of harmful nutrients, and "Nutrients to Encourage" (NTE) models, which emphasize increasing beneficial nutrients. Some modern systems attempt to integrate both approaches for a more balanced assessment. These contrasting philosophies inform public health policies, front-of-pack labeling (FOPL) systems, and agricultural research priorities, particularly in the context of evaluating traditional versus modern crop varieties [11] [14].

The fundamental distinction between these approaches lies in their primary focus. NTL models identify foods high in saturated fat, sodium, and added sugars to help consumers limit these components, while NTE models highlight foods rich in protein, fiber, vitamins, and minerals to encourage their consumption [14] [15]. This philosophical divergence creates tangible differences in how food products are scored, labeled, and ultimately perceived by consumers and researchers.

Comparative Analysis of Major Nutrient Profiling Models

Key Characteristics of Prominent NP Models

Table 1: Comparison of Major Nutrient Profiling Models and Their Philosophical Approaches

Model Name	Primary Philosophy	Nutrients to Limit	Nutrients to Encourage	Food Categorization	Scoring Output
Health Canada FOP Symbol (HC-FOPS) [14]	Nutrients to Limit	Saturated fat, sugars, sodium	None	Not specified	Binary (trigger-based)
PAHO Model [11]	Nutrients to Limit	Saturated fat, trans fat, sodium, free sugars	None	5 categories	Dichotomous ("excessive" or not)
Nutri-Score [14]	Balanced	Saturated fat, sodium, sugars, energy	Protein, fiber, fruits/vegetables/nuts/legumes (FVNL)	2 categories	Ordinal (A to E classes)
Nutrient Rich Foods (NRF) Index [15]	Nutrients to Encourage	Saturated fat, added sugar, sodium	Protein, fiber, vitamin A, C, E, calcium, iron, magnesium, potassium	Not specified	Continuous score
Food Compass 2.0 [5]	Comprehensive Balanced	Saturated fat, sodium, added sugar, processing attributes	Protein, fiber, whole fruits, vegetables, legumes, specific vitamins/minerals, nutrient ratios	Uniform scoring across categories	Continuous (0-100 scale)

Quantitative Comparison of Model Components

Table 2: Detailed Nutrient Components Across Profiling Models

Nutrient/Component	Ofcom	FSANZ	Nutri-Score	HCST	EURO	PAHO	Food Compass 2.0	NRF Index
Saturated Fat	✓	✓	✓	✓	✓	✓	✓	✓
Sodium	✓	✓	✓	✓	✓	✓	✓	✓
Total Sugars	✓	✓	✓		✓		✓
Added/Free Sugars						✓	✓	✓
Energy	✓	✓	✓	✓			✓
Trans Fat	✓	✓				✓	✓
Protein			✓		✓		✓	✓
Fiber			✓		✓		✓	✓
Fruits/Vegetables			✓				✓
Vitamins/Minerals							✓	✓
Food Processing							✓
Nutrient Ratios							✓

Experimental Protocols for NP Model Application

Protocol 1: Comparative Analysis of Traditional vs. Modern Crop Varieties

Objective: To evaluate and compare the nutritional quality of traditional and modern crop varieties using different nutrient profiling models.

Materials and Reagents:

• Traditional and modern cultivars of target crop species
• Laboratory equipment for nutritional analysis (HPLC, GC-MS, ICP-MS)
• Food composition database (e.g., USDA FNDDS, BLS)
• Statistical analysis software (R, SPSS, SAS)

Methodology:

• Sample Preparation: Obtain representative samples of traditional and modern crop varieties from certified sources. Ensure consistent growing conditions, harvest timing, and post-harvest handling to minimize environmental variability.
• Nutritional Analysis: Conduct comprehensive compositional analysis according to standardized methods (AOAC International):
  • Proximate analysis (protein, fat, moisture, ash)
  • Dietary fiber content (enzymatic-gravimetric method)
  • Vitamin and mineral profiles (chromatography, spectrometry)
  • Sugar profiles (HPLC)
  • Fatty acid composition (GC-MS)
  • Phytochemical content (specific to crop type)
• Data Normalization: Convert all nutrient values to standardized reference amounts (typically per 100g or 100kcal edible portion).
• Model Application: Apply selected NP models to the analytical data:
  • Calculate NRF scores according to Drewnowski et al. [15]
  • Compute Nutri-Score values using the published algorithm [14]
  • Determine Food Compass 2.0 scores across all domains [5]
  • Assess HC-FOPS triggering nutrients [14]
• Statistical Analysis: Perform comparative statistical analysis (ANOVA, t-tests) to identify significant differences between traditional and modern varieties under each model.

Expected Outputs:

• Quantitative scores for each variety under different NP models
• Identification of which varieties are favored under NTL vs. NTE approaches
• Correlation analysis between model scores and specific nutrient components

Protocol 2: Validation Against Health Outcome Biomarkers

Objective: To validate NP model scores against clinical biomarkers of health status.

Materials and Reagents:

• Human subjects or appropriate animal models
• Clinical laboratory equipment for biomarker analysis
• Dietary assessment tools (24-hour recalls, FFQ)
• Bio-specimen collection kits (blood, urine)

Methodology:

• Study Design: Implement controlled feeding trial or rigorous observational study with comprehensive dietary assessment.
• Biomarker Assessment: Measure relevant health biomarkers:
  • Blood lipids (LDL-C, HDL-C, triglycerides)
  • Inflammatory markers (CRP, IL-6)
  • Glycemic control markers (fasting glucose, HbA1c)
  • Blood pressure
  • Body composition
• Diet Scoring: Calculate individual NP scores based on consumed diets:
  • Compute energy-weighted average Food Compass score (i.FCS) [5]
  • Determine overall Nutri-Score pattern [14]
  • Calculate mean NRF index [15]
• Association Analysis: Conduct multivariate regression analysis to examine relationships between NP model scores and health biomarkers, adjusting for confounding factors (age, sex, BMI, physical activity).

Validation Metrics:

• Strength of association (regression coefficients, odds ratios)
• Predictive accuracy (ROC curves, sensitivity/specificity)
• Dose-response relationships

Research Workflow and Conceptual Framework

The Researcher's Toolkit: Essential Reagents and Materials

Table 3: Essential Research Reagents and Materials for Nutrient Profiling Studies

Category	Specific Items	Application/Function	Example Models
Analytical Standards	Certified reference materials (CRMs) for nutrients, Sugar standards, Fatty acid methyl esters, Vitamin standards, Mineral standards	Instrument calibration and method validation	All models
Chromatography Supplies	HPLC columns (C18, HILIC), GC columns, SPE cartridges, Mobile phase reagents	Separation and quantification of specific nutrients	Food Compass, NRF, Nutri-Score
Spectroscopy Supplies	ICP-MS calibration standards, AAS lamps, UV-Vis cuvettes	Elemental analysis and concentration determination	NRF, Food Compass
Software Tools	Statistical packages (R, SAS), Database management systems, Algorithm implementation code	Data analysis and model calculation	All models
Laboratory Equipment	Freeze dryer, Analytical balance, Microwave digestion system, Automated extraction	Sample preparation and processing	All models
Bioinformatics Resources	Food composition databases, Metabolomics libraries, Genome databases for traditional varieties	Data integration and interpretation	Food Compass, NRF

Application Notes and Technical Considerations

Model Selection Criteria for Traditional vs. Modern Crop Research

When evaluating traditional and modern crop varieties, the choice of NP model significantly influences research outcomes and conclusions. NTL-focused models like HC-FOPS and PAHO are particularly effective for identifying varieties with reduced levels of potentially harmful components, making them suitable for research targeting specific dietary restrictions or public health initiatives aimed at reducing chronic disease risk factors [11] [14]. However, these models may overlook the potential benefits of traditional varieties that contain beneficial phytochemicals or complementary nutrient profiles.

NTE-focused models such as the NRF Index provide advantages for identifying nutrient-dense traditional varieties that may offer superior nutritional quality despite potentially higher levels of some limiting nutrients [15]. This approach aligns with research objectives focused on addressing micronutrient deficiencies or maximizing nutritional yield per cultivation area.

Comprehensive balanced models including Food Compass 2.0 and Nutri-Score offer the most holistic assessment for comparing traditional and modern varieties, as they capture both positive and negative aspects of nutritional composition [14] [5]. Food Compass 2.0's additional consideration of food processing characteristics and ingredient quality makes it particularly suitable for evaluating the full spectrum of nutritional differences between traditional landraces and highly processed modern food products.

Implementation Challenges and Methodological Considerations

Researchers must address several technical challenges when applying NP models to traditional vs. modern crop comparisons:

Data Quality and Completeness: Many traditional varieties have incomplete nutritional data in existing databases, requiring primary compositional analysis. Standardized analytical methods must be employed to ensure comparability across varieties [3].

Reference Amount Standardization: Different models use different reference amounts (per 100g, per 100kcal, per serving), which can significantly affect scores, especially for crops with varying water content or energy density [11] [5].

Bioavailability Considerations: Traditional varieties may contain antinutritional factors or have different nutrient bioavailability than modern varieties, which NP models typically do not account for [3].

Cultural and Dietary Context: The relevance of specific nutrients may vary across different populations and dietary patterns, particularly when studying traditional crops consumed in specific cultural contexts [13].

Future Directions and Model Evolution

The field of nutrient profiling continues to evolve with several emerging trends relevant to traditional vs. modern crop research:

Dynamic Nutrient Profiling: Advanced systems now incorporate real-time data, biomarker integration, and artificial intelligence to provide more personalized assessments [16] [17]. These approaches could be adapted to account for individual variations in nutrient metabolism and health status when evaluating crop varieties.

Integration of Sustainability Metrics: Future NP models may incorporate environmental impact assessments alongside nutritional quality, creating a more comprehensive evaluation framework for crop improvement programs [3].

Multi-omics Integration: The incorporation of genomic, proteomic, and metabolomic data into nutrient profiling could provide deeper insights into the genetic determinants of nutritional quality in traditional and modern varieties [17] [5].

Traditional Crop-Specific Adaptations: As research accumulates on the unique nutritional properties of traditional varieties, model adjustments may be warranted to better capture their potential health benefits, such as specific phytochemical profiles or synergistic nutrient interactions [3].

Nutrient Profiling Models (NPMs) are quantitative algorithms designed to evaluate and rank the healthfulness of foods and beverages based on their nutritional composition and other health-relevant characteristics [5]. These scientific tools are increasingly deployed by governments and industry to support public health initiatives, including front-of-package labelling (FOPL), food reformulation targets, restrictions on food marketing to children, and guidance for health-conscious investment strategies [5] [18]. The escalating global health burden of diet-related non-communicable diseases has accelerated the need for validated, science-based systems that can empower consumers to make healthier choices and encourage manufacturers to improve the nutritional quality of the food supply [19] [20].

This overview focuses on three prominent global NPM frameworks: the Nutri-Score (NS), Health Star Rating (HSR), and Food Compass (FC). While NS and HSR are widely implemented as front-of-pack labelling systems in several countries, Food Compass represents a more recent, comprehensive profiling system developed to address limitations in existing models [18] [21]. Understanding their distinct algorithms, applications, and validation is crucial for researchers investigating the health impacts of traditional versus modern food varieties, as these models provide standardized methodologies for quantifying and comparing their nutritional value.

Model Frameworks and Scoring Algorithms

Nutri-Score (NS)

Nutri-Score is a five-tier, colour-coded front-of-pack grading scheme widely adopted in several European countries [19]. Its algorithm was built upon the British Food Standards Agency nutrient profiling model (FSA-NPM) and is designed to allow consumers to quickly compare the healthfulness of similar food products [19]. The system summarizes a product's overall nutritional quality by assigning a letter from A (dark green, representing the healthiest option) to E (dark orange, representing the least healthy option) [19].

Key Algorithmic Characteristics:

• Profiling Categories: It classifies products into four main categories for scoring: 1) Beverages, 2) Foods (general), 3) Added fats, and 4) Cheese [19].
• Negative Points (0-10 points each): Assessed for energy (kcal), saturated fat, total sugars, and sodium content per 100g or 100ml [19].
• Positive Points (0-5 points each): Awarded for the content of fruits, vegetables, nuts, legumes, as well as fiber and protein [19].
• Specific Provisions: The system automatically assigns a grade 'A' to plain water and includes special considerations for certain oils (rapeseed, walnut, olive oil) within the fruit/vegetable/nut/legume component [19].

Health Star Rating (HSR)

The Health Star Rating is a voluntary front-of-pack labelling system developed by the Australian government in collaboration with industry and consumer groups [19]. It presents an overall healthfulness rating on a scale from half a star (least healthy) to five stars (most healthy) in half-star increments [19] [22]. Like Nutri-Score, the HSR algorithm is derived from the Ofcom model but incorporates extended score scales for several nutrients.

Key Algorithmic Characteristics:

• Profiling Categories: Products are categorized as 1) Non-dairy beverages, 2) Foods (general), 3) Oils and spreads, 1D) Dairy beverages, 2D) Dairy foods, or 3D) Cheese, with specific algorithms for each [19].
• Baseline Points (0-30 points for some): Awarded for undesirable components, with varying maximum points depending on the nutrient and category (e.g., saturated fat and sodium can contribute up to 30 baseline points in some categories, compared to 10 in NS) [19].
• Modifying Points (0-15 points for some): Awarded for desirable components like protein, fiber, and the content of fruits, vegetables, nuts, and legumes (FVNL), with higher maximum points for protein (15), fiber (15), and FVNL (up to 10) than in NS [19].
• Automatic Ratings: Plain water automatically receives 5 stars, unsweetened flavoured water receives 4.5 stars, and fresh/minimally processed fruits and vegetables receive 5 stars [19].

Food Compass (FC)

Food Compass is a comprehensive nutrient profiling system developed by researchers at Tufts University to address perceived limitations in existing models [18] [21]. Its most recent version, Food Compass 2.0, incorporates the latest scientific evidence on diet-health relationships, including food processing, specific ingredients, and additives [5] [20]. It generates a single score from 1 (least healthy) to 100 (most healthy) for any food, beverage, or even mixed meal, using a uniform set of criteria across all categories [22] [23].

Key Algorithmic Characteristics:

• Scoring Basis: Unique among the three, it scores foods per 100 kcal (418.4 kJ) to facilitate comparison of items differing greatly in bulk and to enable scoring of mixed dishes and meals [23].
• Multi-Domain Assessment: It evaluates 54 attributes across 9 distinct health-relevant domains [23] [18]:
  • Nutrient Ratios (e.g., unsaturated:saturated fat ratio)
  • Vitamins
  • Minerals
  • Food-based Ingredients
  • Additives
  • Processing (e.g., NOVA classification, fermentation)
  • Specific Lipids (e.g., omega-3 fatty acids)
  • Fiber & Protein
  • Phytochemicals (e.g., flavonoids, carotenoids)
• Scoring Method: Beneficial attributes are scored from 0 to 10, harmful attributes from -10 to 0, and ratios (which can be harmful or beneficial) from -10 to 10. Domain scores are averaged or summed, then all domain scores are combined with specific weights (full weight for the first six domains, half weight for the last three) to produce the final score [23].
• Healthfulness Cut-Points: A score ≥70 identifies foods/beverages to be encouraged; 31-69, consumed in moderation; and ≤30, minimized [22].

Table 1: Comparative Overview of NPM Frameworks

Feature	Nutri-Score	Health Star Rating	Food Compass
Origin/Region	France/Europe	Australia/New Zealand	Tufts University (Global)
Graphic Format	5-colour scale (A to E)	10-point monochrome star scale (0.5 to 5)	Numerical score (1 to 100)
Scoring Basis	Per 100g / 100ml	Per 100g / 100ml	Per 100 kcal
Number of Attributes	Limited set (~7-8)	Limited set (~7-8)	54 attributes across 9 domains
Includes Processing	No	No	Yes (NOVA, fermentation, frying)
Includes Additives	No	No	Yes
Includes Phytochemicals	No	No	Yes
Primary Use Case	Front-of-Pack Labelling	Front-of-Pack Labelling	Research, Policy, Product Reformulation

Quantitative Data and Comparative Performance

Direct Model Comparisons

Studies have directly investigated the alignment and correlation between these profiling systems. A 2023 study profiling over 17,000 products from the Slovenian food supply found that NS and HSR, given their common ancestral algorithm (Ofcom), were largely aligned. The overall agreement between them was 70% (Cohen’s κ = 0.62), with a very strong correlation (Spearman's rho = 0.87) [19]. However, notable disagreements emerged in specific categories:

• Cheeses: Agreement was very poor (8%, κ = 0.01), with HSR classifying most cheeses (63%) as healthy (≥3.5 stars), while NS predominantly assigned lower scores [19].
• Cooking Oils: Agreement was low (27%, κ = 0.11). NS favored olive and walnut oil, while HSR favored grapeseed, flaxseed, and sunflower oil [19].

In contrast, Food Compass demonstrates a more moderate correlation with both HSR and NS. One analysis reported an overall correlation of Spearman r=0.67 between the original Food Compass and HSR, meaning HSR explained only about 45% of the variation in Food Compass scores [22]. This correlation varied widely by food category, being much lower for grains, vegetables, legumes, nuts, seeds, and savory snacks and sweet desserts [22]. The very high correlation between NS and HSR (often r≈0.90) indicates they provide similar information, whereas Food Compass captures distinct and additional dimensions of healthfulness [5] [22].

Validation Against Health Outcomes

A key measure of an NPM's utility is its ability to predict health outcomes when applied to population-level dietary intake.

• Food Compass: In a nationally representative sample of 47,099 U.S. adults, each 10.8-point (1 standard deviation) higher energy-weighted average individual Food Compass Score (i.FCS) was associated with multiple improved health parameters after multivariable adjustment [5]. These included lower BMI (-0.56 kg/m²), systolic blood pressure (-0.55 mm Hg), LDL cholesterol (-1.49 mg/dL), HbA1c (-0.02%), and lower prevalence of metabolic syndrome, cardiovascular disease, and cancer [5]. A higher i.FCS was also associated with an 8% lower risk of all-cause mortality per 1 s.d. increase [5] [20].
• Nutri-Score & Health Star Rating: These systems have also demonstrated validity in predicting health outcomes. A 2024 study comparing NPMs found that all were significantly associated with healthier weight and blood pressure measures, though the strength of associations varied [24]. Diets rated healthier by these models were consistently linked to better cardiometabolic profiles.

Table 2: Association of NPMs with Health Outcomes in Observational Studies

Health Outcome	Nutri-Score	Health Star Rating	Food Compass
Obesity (BMI)	Significant inverse association [24]	Significant inverse association [24]	-0.56 kg/m² per 1 s.d. [5]
Systolic Blood Pressure	Significant inverse association [24]	Significant inverse association [24]	-0.55 mm Hg per 1 s.d. [5]
LDL Cholesterol	Data from specific studies	Data from specific studies	-1.49 mg/dL per 1 s.d. [5]
Metabolic Syndrome Prevalence	Data from specific studies	Data from specific studies	Lower prevalence (OR 0.86) [5]
All-Cause Mortality	Associated with lower risk [24]	Associated with lower risk [24]	24% lower risk in highest vs. lowest quintile [5]

Application Notes and Experimental Protocols

Protocol for Applying NPMs to Food Composition Data

This protocol outlines the steps for researchers to calculate the Nutri-Score, Health Star Rating, and Food Compass for a given food product using compositional data.

1. Data Collection and Preparation:

• Input Data: Gather detailed quantitative data on the food's composition per 100g or 100ml, or per 100 kcal for Food Compass. Required data includes:
  • Macronutrients: Energy (kcal/kJ), protein (g), total fat (g), saturated fat (g), carbohydrates (g), total sugars (g), fiber (g), sodium (mg).
  • Micronutrients: Vitamins (A, C, D, B vitamins, etc.), minerals (calcium, iron, potassium, etc.).
  • Food Components: Percentages of fruits, vegetables, nuts, legumes (FVNL), whole grains.
  • Other Attributes (Critical for Food Compass): Presence of additives (sweeteners, colors, nitrites), processing level (NOVA classification), specific fatty acids (ALA, EPA+DHA), and phytochemical content (if available) [23].

2. Categorization:

• Correctly assign the food to its specific category within each NPM's classification system, as algorithms differ [19] [23].
  • For Nutri-Score: Determine if the item is a "Beverage," "Food," "Added Fat," or "Cheese."
  • For HSR: Determine if it is a "Non-dairy Beverage," "Food," "Oil/Spread," "Dairy Beverage," "Dairy Food," or "Cheese."
  • For Food Compass: While it uses one uniform algorithm, note the food group (e.g., vegetable, meat, grain) for later interpretation.

3. Score Calculation:

• Nutri-Score/HSR: Apply the respective algorithm [19].
  • Calculate "Negative" (baseline) points for energy, saturated fat, sugars, sodium.
  • Calculate "Positive" (modifying) points for FVNL, protein, fiber.
  • For HSR, subtract positive points from baseline points to get the final score, which is then converted to a star rating.
  • For Nutri-Score, the net score is translated into the A-E scale.
• Food Compass: For each of the 54 attributes, calculate a score based on its content per 100 kcal [23].
  • Score beneficial attributes 0-10, harmful attributes -10 to 0, and ratios -10 to 10.
  • Calculate the average score for each of the 9 domains.
  • Sum the domain scores (with equal weight for domains 1-6 and half weight for domains 7-9).
  • Scale the final sum to a value between 1 and 100.

4. Validation and Quality Control:

• Where possible, cross-validate calculated scores for a subset of products against scores published in official databases or previous studies [19] [22].
• For Food Compass, ensure consistent application of the NOVA processing classification system across all products, using standardized definitions [23].

Research Reagent Solutions: Essential Materials for NPM Analysis

Table 3: Key Reagents and Tools for NPM Research

Item/Tool	Function in NPM Research	Example Sources/Notes
Food Composition Database	Provides standardized nutrient data for a wide range of foods. Essential for calculating scores.	USDA FNDDS [23], FoodSwitch [23], Branded Food Products Database (CLAS) [19]
NOVA Classification Guide	Provides rules for categorizing foods by degree of processing, a critical input for Food Compass.	Monteiro et al. (2019) reference guidelines [23]
Food Processing Data	Data on additives, fermentation, and frying, required for specific domains in Food Compass.	Ingredient lists; specialized databases on food additives [5] [23]
Phytochemical Database	Provides data on flavonoid and carotenoid content for scoring the Phytochemicals domain in Food Compass.	USDA Flavonoid Database [23], Phenol-Explorer
Standardized NPM Algorithm Scripts	Software code (e.g., in R or Python) that automates the complex calculations for each NPM.	Custom scripts based on published algorithms [19] [23]; Tufts Food Compass website [18]

Visualizations and Workflows

NPM Algorithm Selection and Application Workflow

The following diagram illustrates the decision-making workflow for selecting and applying an appropriate NPM based on research objectives.

Food Compass Multi-Domain Scoring System

This diagram depicts the complex, multi-domain structure of the Food Compass 2.0 scoring algorithm.

Application Note: Characterizing Traditional Food Healthfulness

Background and Rationale

Traditional foods present a unique challenge for modern nutrient profiling systems (NPS), which were predominantly developed to address public health issues in Western high-income countries—specifically, the overconsumption of calories, saturated fats, sugars, and salt [25]. These systems often employ a reductionist approach, evaluating foods based on isolated nutrients rather than holistic dietary patterns or cultural context. In contrast, traditional foods are deeply embedded in cultural heritage and often represent centuries of accumulated wisdom about food combinations, preparation methods, and seasonal eating patterns that promote health [26]. The global shift toward diets centered on limited crops has marginalised many traditional foods, potentially diminishing nutritional diversity and culinary heritage [27].

Nutrient profiling models like Health Star Rating, Nutri-Score, and the recently updated Food Compass 2.0 aim to quantitatively assess the healthfulness of foods [5]. However, these systems may inadvertently penalize traditional foods that:

• Contain naturally occurring sugars and fats within a whole-food matrix
• Utilize traditional preparation methods like fermentation, soaking, or sprouting
• Represent culturally significant dietary patterns with demonstrated health benefits
• Provide essential micronutrients that are deficient in low- and middle-income countries (LMICs) [25]

This application note provides structured methodologies to bridge this gap, enabling researchers to evaluate traditional foods through both standardized metrics and contextually appropriate holistic frameworks.

Key Analytical Challenges

Research indicates several specific challenges in applying standardized NPS to traditional foods:

• Processing Considerations: Modern NPS like Food Compass 2.0 now provide positive points for non-ultraprocessed foods rather than only penalizing ultraprocessed foods, better accommodating traditionally processed foods like fermented items [5]. However, traditional processing methods like nixtamalization of corn (which increases niacin bioavailability) may not be fully captured [26].

• Nutrient Synergy: Traditional food combinations (e.g., rice and lentils forming complete proteins) demonstrate nutritional benefits that reductionist approaches may miss by focusing on individual nutrients [26].

• Cultural Context: In Zimbabwe, traditional food consumption is facilitated by generational factors and family influence but hindered by perceptions of inconvenience and aggressive marketing of processed foods [28].

• Diverse Health Priorities: While high-income countries focus on limiting "negative" nutrients, LMICs continue to face challenges of undernutrition and micronutrient deficiencies, requiring NPS that emphasize adequate intake of vitamins, minerals, and high-quality protein [25].

Table 1: Comparative Analysis of Nutrient Profiling Systems for Traditional Food Evaluation

Profiling System	Primary Focus	Strengths for Traditional Foods	Limitations for Traditional Foods
Food Compass 2.0 [5]	Comprehensive healthfulness across 9 domains	Uniform scoring for mixed foods/meals; considers processing	May undervalue traditional preparation methods
Nutri-Score [25]	Nutrients to limit (energy, sugar, fat, salt)	Simple consumer communication	Penalizes high-energy traditional foods like nuts
Health Star Rating [5]	Basic nutrient balance	Category-specific comparisons	Inconsistent scoring across food categories
Traditional Food Systems [26]	Holistic dietary patterns	Captures food synergies and cultural context	Difficult to standardize and quantify

Experimental Protocols

Protocol 1: Comprehensive Traditional Food Nutrient Analysis

Purpose and Scope

To systematically evaluate the nutritional composition of traditional foods using both standardized nutrient profiling and contextually appropriate adaptations that account for traditional preparation methods, cultural significance, and local health priorities. This protocol is particularly relevant for researchers investigating traditional foods from diverse cultural backgrounds, including indigenous food systems and ethnic cuisines.

Equipment and Materials

Table 2: Research Reagent Solutions for Traditional Food Analysis

Item	Function	Application Notes
Food Composition Databases (e.g., INFOODS, USDA SR-28) [25]	Provides baseline nutrient data	Critical for establishing reference values; regional databases preferred
Traditional Food Knowledge Holders [26]	Contextual understanding of preparation and use	Essential for culturally appropriate research design
Statistical Analysis Software (R, Python, SPSS) [29]	Data analysis and pattern recognition	Enables application of clustering, correlation, and dimension reduction techniques
Laboratory Equipment for Chemical Analysis [29]	Verification of nutrient composition	Validates database information against actual food samples
Cultural Validation Framework [26]	Assesses cultural appropriateness	Ensures research respects traditional knowledge systems

Step-by-Step Procedure

• Food Identification and Selection

  • Identify traditional foods through ethnographic methods including community engagement and consultation with cultural knowledge holders [26]
  • Document traditional preparation methods, seasonal variations, and consumption patterns
  • Select representative samples for analysis, noting geographic and seasonal variations

• Nutrient Composition Analysis

  • Obtain nutrient composition data from validated databases (e.g., INFOODS, USDA SR-28) [25]
  • For missing data, conduct laboratory analysis for key nutrients based on local nutritional priorities (e.g., iron, zinc, vitamin A in LMICs) [8]
  • Record values per 100g and per 100kcal to accommodate different profiling systems

• Application of Multiple Profiling Systems

  • Calculate scores using standardized NPS (Food Compass 2.0, Nutri-Score, HSR) [5]
  • Adapt scoring to emphasize regionally relevant nutrients of concern [25]
  • Document discrepancies between different systems and potential misclassification of traditional foods

• Holistic Factor Assessment

  • Evaluate processing characteristics using the NOVA system, noting traditional vs. industrial processing methods [5]
  • Document food synergies in traditional combinations (e.g., rice + lentils for complete protein) [26]
  • Assess cultural significance and consumption patterns through community surveys [28]

• Data Analysis and Validation

  • Apply statistical methods including principal component analysis, cluster analysis, and correlation studies to identify nutrient patterns [29]
  • Validate profiling results against health outcomes where population data exists [5]
  • Compare traditional diet patterns with modern patterns for key health parameters [13]

Protocol 2: Traditional Food System Health Outcome Validation

Purpose and Scope

To validate the health impacts of traditional dietary patterns through epidemiological studies and clinical interventions, bridging traditional knowledge with evidence-based nutrition science. This protocol is designed for research teams working with population health data or conducting dietary intervention studies, particularly in communities undergoing nutrition transition.

Equipment and Materials

• Population health datasets with dietary intake records
• Statistical software for multivariate analysis
• Traditional food frequency questionnaires, culturally adapted
• Anthropometric measurement equipment
• Biomarker analysis capabilities (blood glucose, lipids, inflammatory markers)

Step-by-Step Procedure

• Study Population Selection

  • Identify populations with varying adherence to traditional food patterns
  • Include sufficient sample size to detect significant health outcome differences
  • Stratify by age, gender, and other relevant demographic factors

• Dietary Pattern Assessment

  • Develop and administer traditional food frequency questionnaires [13]
  • Calculate individual Food Compass Scores (i.FCS) based on energy-weighted average of foods consumed [5]
  • Classify participants according to dietary pattern adherence (traditional, western, health-conscious) [13]

• Health Outcome Measurement

  • Collect anthropometric data (BMI, waist circumference, body composition) [13]
  • Measure clinical biomarkers (blood lipids, glucose, blood pressure) [5]
  • Document prevalence of health conditions (metabolic syndrome, CVD, cancer) [5]

• Statistical Analysis

  • Conduct multivariate adjustment for confounding variables (age, gender, physical activity, socioeconomic status)
  • Calculate odds ratios for health conditions across dietary pattern groups
  • Determine hazard ratios for all-cause mortality by dietary pattern [5]

• Traditional-Modern Diet Comparison

  • Compare nutrient profiles between traditional and modern dietary patterns [13]
  • Analyze differences in saturated fat, PUFA, and dietary fiber intake between patterns [13]
  • Correlate traditional food consumption with favorable health outcomes

Data Analysis and Interpretation Framework

Quantitative Assessment Metrics

Table 3: Key Metrics for Traditional Food Pattern Analysis Based on Validation Studies

Metric Category	Specific Measures	Traditional Diet Findings	Modern Diet Findings
Anthropometric [13]	BMI, body fat percentage	Lower likelihood of overweight/obesity	Higher BMI and body fat percentage
Cardiometabolic [5]	Blood pressure, LDL cholesterol, HDL cholesterol	Improved lipid profiles, lower blood pressure	Less favorable cardiometabolic profiles
Disease Prevalence [5]	Metabolic syndrome, CVD, cancer	Lower prevalence rates	Higher prevalence rates
Nutrient Intake [13]	SFA, PUFA, dietary fiber	Higher SFA, lower PUFA and fiber in some traditional patterns	Varies by pattern type
Mortality [5]	All-cause mortality hazard ratio	24% lower risk in highest vs. lowest i.FCS quintile	Higher mortality risk with lower quality diets

Interpretation Guidelines

When analyzing traditional foods through nutrient profiling systems, researchers should consider:

• Contextualizing "Negative" Nutrients: Saturated fats in traditional diets may come from whole food sources like dairy and meat with different metabolic effects than processed equivalents [30]. Food Compass 2.0 has been updated to reflect research suggesting relatively neutral health effects of dairy fat [5].

• Processing Context: Traditional processing methods like fermentation may enhance nutritional value (e.g., increasing bioavailability) rather than degrade it [26].

• Dietary Pattern Synergy: Evaluate traditional foods within their consumption context, as combinations may provide nutritional benefits not apparent when assessing individual items [26].

• Regional Nutritional Priorities: In LMICs, traditional foods may provide critical micronutrients (iron, zinc, vitamin A) that should be weighted more heavily than in high-income countries [25] [8].

The integration of traditional food knowledge with modern nutrient profiling requires a balanced approach that respects cultural context while applying scientific rigor. The methodologies presented here enable researchers to:

• Systematically evaluate traditional foods using both standardized and adapted metrics
• Validate the health impacts of traditional dietary patterns through epidemiological methods
• Develop more nuanced nutrient profiling systems that account for traditional preparation methods and food combinations

Future research should focus on building comprehensive databases of traditional food composition, developing validated metrics for food synergy, and creating flexible profiling systems that can be adapted to regional nutritional priorities. This integrated approach promises to preserve nutritional diversity while advancing evidence-based nutrition science.

Methodologies for Profiling Diverse Food Systems: From Algorithm Design to Real-World Application

Nutrient Profiling Models (NPMs) are scientific tools that classify foods based on their nutritional composition to support public health policies. Within research on traditional versus modern food varieties, these models provide an objective methodology to quantify and compare their nutritional value. This document deconstructs the scoring algorithms of established NPMs, detailing their operational protocols for application in academic and industrial research, including drug and supplement development.

The Core Scoring Algorithms of Major Nutrient Profiling Models

The following section delineates the foundational scoring mechanisms of two pivotal models: the UK's 2004-2005 NPM and its proposed 2018 successor.

The UK 2004-2005 Nutrient Profiling Model (2004 NPM)

The 2004 NPM employs a points-based system applied to 100g of a food or beverage product [31] [32]. Points ('A' points) are awarded for nutrients to limit: energy, saturated fat, total sugars, and sodium. Conversely, points ('C' points) are awarded for beneficial constituents: fruit, vegetable, and nut (FVN) content, dietary fibre, and protein. The final score is the difference between 'A' and 'C' points [31].

Final Score Calculation: A Points (Nutrients to Limit) - C Points (Beneficial Nutrients)

A food product is classified as High in Fat, Salt, or Sugar (HFSS) if it scores 4 or more points. A drink is classified as HFSS if it scores 1 or more point [31] [32]. This model underpins HFSS regulations in Wales and England, governing advertising and promotion restrictions [32].

The Proposed 2018 Updated NPM

Developed by Public Health England, the 2018 NPM retains the basic structure of the 2004 model but introduces key modifications to align with contemporary dietary guidance [31]. The scoring framework remains a points system per 100g, but with critical adjustments to the nutrients assessed.

Table 1: Key Differences Between the 2004 and 2018 UK Nutrient Profiling Models [31]

Feature	2004 NPM	2018 NPM
Sugars Metric	Total Sugars	Free Sugars
Fibre Metric	AOAC fibre	UK definition of fibre (includes all non-starch polysaccharides)
Points for FVN	Awarded for % fruit, vegetable, and nut content	Awarded for % fruit and vegetable content only (nuts removed)
Drinks Criteria	Drinks score ≥1 are HFSS	Drinks score ≥1 are HFSS (but criteria are stricter)
Juices	Treated similarly to other drinks	Classified as foods, not drinks

A primary driver for the update was the Scientific Advisory Committee on Nutrition's (SACN) 2015 report, which recommended reducing free sugars intake and increasing dietary fibre [31]. The shift to free sugars presents a significant technical challenge, as this data is not typically available on standard nutrition labels and requires estimation from ingredient lists [31].

Table 2: Impact of Model Change on Product Categorization [31]

Product Category	% of Products Passing 2004 NPM	% of Products Passing 2018 NPM	Change (Percentage Points)
Beverages	~40%	~10%	-75%
Breakfast Cereals	~85%	~74%	-11%
Yoghurts	~80%	~75%	-5%
Frozen Foods	~80%	~74%	-6%
Cakes	~15%	~18%	+3%

Experimental Protocol for Determining NPM Score

This protocol outlines the methodology for calculating the HFSS status of a food product using the UK 2004-2005 NPM, suitable for research on both traditional and modern food formulations.

Research Reagent Solutions and Essential Materials

Table 3: Key Materials for NPM Score Determination

Item	Function/Description
Product Sample	A homogeneous, representative sample of the food or drink product to be analyzed.
Nutritional Composition Data	Accurate quantitative data for energy (kcal/kJ), saturated fat, total sugars, sodium, protein, and dietary fibre per 100g/ml. Must be derived from laboratory analysis or reliable database.
Ingredient Declaration	A complete list of ingredients, required to determine the percentage of fruit, vegetable, and nut (FVN) content.
2004 NPM Technical Guidance	The official scoring tables and rules for assigning 'A' and 'C' points [32].
McCance and Widdowson's Composition of Foods	A recognized UK database for "generally established and accepted data" where specific product data is unavailable [32].
Standardized Calculation Sheet	A software tool or spreadsheet pre-programmed with the NPM algorithm to ensure scoring accuracy and reproducibility.

Step-by-Step Workflow

• Sample Preparation and Data Acquisition

  • For pre-packaged goods, obtain the nutritional information from the product label or manufacturer's specifications.
  • For novel food varieties or raw ingredients, conduct standardized laboratory nutritional analysis to obtain values per 100g for all 'A' and 'C' nutrients.
  • Secure the full ingredient list to calculate the percentage of FVN content.

• Assign 'A' Points (Nutrients to Limit)

  • Refer to the official NPM points table.
  • For each nutrient (energy, saturated fat, total sugars, sodium), determine the points based on its content per 100g/ml.
  • Record the points for each nutrient and sum them for the total 'A' score.

• Assign 'C' Points (Beneficial Nutrients)

  • Determine the percentage content of FVN. Assign points based on the percentage band (e.g., 40-60%, >60%) [31].
  • Assign points for fibre content per 100g based on defined bands.
  • Assign points for protein content per 100g based on defined bands.
  • Record the points for each component and sum them for the total 'C' score.

• Final Score Calculation and Classification

  • Calculate the final NPM score: Total A Points - Total C Points.
  • Classify the product:
    • For Foods: A score of 4 or above classifies the product as HFSS.
    • For Drinks: A score of 1 or above classifies the product as HFSS.

Special Considerations in Experimental Design

• Product State: The NPM score is typically calculated for the product "as sold." For products requiring reconstitution (e.g., powdered drinks), the score must be based on 100g of the product as reconstituted according to manufacturer's instructions [32].
• Composite Products: For multi-component items (e.g., a ready meal with meat and sauce), if nutrition information is provided for the whole product, calculate the score on this basis. If information is separate, calculate the score only for the component falling under a regulated category (e.g., the sauce) [32].
• Data Integrity: Researchers must document the source of all nutritional data. When using estimated or database values, this should be clearly stated to ensure the reproducibility of the HFSS classification.

Model Evolution and Validation

The transition from the 2004 NPM to updated models like the 2018 NPM or the 2023 Nutri-Score NPM represents an evolution in scientific understanding. A 2025 study validated the updated Nutri-Score model in a French population, demonstrating that the 2023 NPM Dietary Index (DI) was more strongly correlated with the consumption of healthy fats and whole grains, and less strongly correlated with cheeses and refined cereals, compared to its 2015 predecessor [33]. This indicates that modernized models better reflect current dietary guidance. For global contexts, particularly Low- and Middle-Income Countries (LMICs), NP models may need to be reconceptualized to address micronutrient deficiencies and protein quality, rather than solely focusing on nutrients to limit [25].

Visual Representation of NPM Scoring Workflow

The following diagram illustrates the logical sequence for determining a product's HFSS status using the 2004 NPM algorithm.

Visualizing the Historical Evolution of Nutrient Profiling Models

The following diagram maps the key evolutionary milestones and driving factors behind major NPM revisions.

Nutrient profiling models are essential tools for classifying and ranking foods based on their nutritional composition to support public health research, policy, and product development [25]. The accuracy and applicability of these models fundamentally depend on the quality and characteristics of the underlying nutrient databases. Researchers face significant challenges in navigating the distinct landscapes of data sources for traditional whole foods versus modern branded products [34]. This application note provides a structured comparison of these database types, details experimental protocols for their effective utilization, and presents visualization tools to guide researchers in selecting appropriate data sourcing strategies for nutrient profiling studies, particularly those comparing traditional versus modern food varieties.

Database Characteristics and Comparative Analysis

Nutrient composition data is primarily sourced from two complementary types of databases: National Food Composition Databases (FCDBs) for traditional and generic foods, and Branded Product Databases for specific manufactured items [34]. Understanding their respective strengths and limitations is crucial for appropriate research design.

Table 1: Comparative Characteristics of Major Nutrient Database Types

Characteristic	National Food Composition Databases (FCDBs)	Branded Product Databases
Primary Source	Laboratory analysis of commodity foods, scientific literature, and standardized recipes [35] [34]	Product nutrition labels, manufacturer-submitted information [35] [34]
Number of Foods	Limited (e.g., ~3,000 in UK's CoFID) [34]	Extensive (e.g., >200,000 in USDA Branded Foods) [35] [25]
Nutrient Coverage	Comprehensive (e.g., >100 nutrients including micronutrients) [34]	Limited to label mandates (energy, macros, sodium, etc.) [34]
Update Frequency	Infrequent (e.g., every few years) [34]	Frequent (e.g., monthly for USDA Branded Foods) [35]
Data Transparency	High (documented analytical methods/sample sizes) [34]	Low (method of derivation not required on label) [34]
Ideal Application	Assessing micronutrient intake, dietary pattern analysis, epidemiological research [34] [36]	Monitoring product reformulation, assessing market trends, consumer-facing labeling [35] [37]

Key databases for public health research include the USDA's FoodData Central, which integrates multiple data types: Foundation Foods (current analytical data), SR Legacy (historical data), FNDDS (foods as consumed in NHANES), and the Branded Foods module [35] [36]. The Food and Nutrient Database for Dietary Studies (FNDDS) 2021-2023 continues a shift toward more generic food versions while incorporating recent analytical data, which is critical for national dietary surveillance [36].

Table 2: Key Nutrient Databases for Research Use

Database Name	Primary Content	Update Frequency	Notable Features
USDA FoodData Central [35]	Integrated data from multiple sources (analytical, branded, legacy)	Varies by type (e.g., Branded: Monthly; Foundation: Twice yearly)	Single portal for diverse data types; Public domain (CC0 license)
USDA FNDDS [36]	Foods/beverages as consumed in WWEIA, NHANES	Every 2 years	Designed for dietary recall coding; 65 nutrients/food components
USDA Global Branded Food Products [35]	Label data from commercial brands	Monthly	Captures market innovation and reformulation
INFOODS/FAO [25]	International food composition data	Varies by country	Essential for global studies and low/middle-income country contexts

Experimental Protocols for Database Utilization

Protocol 1: Sourcing and Mapping Data for Traditional Food Varieties

Purpose: To acquire and standardize nutrient composition data for traditional, minimally processed, and generic food items for use in nutrient profiling model development.

Materials:

• Primary Database: USDA FoodData Central (Foundation Foods and SR Legacy) [35]
• Supplementary Databases: FAO/INFOODS for international focus, regional databases (e.g., SMILING for Southeast Asia) [25]
• Software: Statistical analysis software (e.g., R, Python) or specialized dietary assessment platforms

Procedure:

• Define Food List: Compile a target list of traditional foods and varieties relevant to the research question (e.g., heritage grains, traditional vegetable cultivars).
• Data Extraction: Query the primary database via its API or web interface using standard food identifiers or keywords. Download data for all available nutrients per 100g edible portion.
• Handle Missing Data: Identify nutrients with missing values. Implement a predefined hierarchy for data imputation: a) values from a closely related food item, b) values from recipe simulation, c) values from scientific literature for the same food [34] [36].
• Standardize Units: Convert all nutrient values to consistent units (e.g., mg, µg) per 100g of food.
• Data Validation: Cross-reference a subset of key nutrients with published scientific literature or other national databases to identify potential outliers.
• Create Research Dataset: Compile the validated, standardized data into a structured dataset (e.g., CSV, SQL database) for model input, documenting all imputation decisions.

Protocol 2: Sourcing and Augmenting Data for Branded Food Products

Purpose: To acquire nutrient data for modern, branded food products and augment label-level data with estimated micronutrient values where necessary.

Materials:

• Primary Database: USDA Branded Food Products Database [35] or commercial branded data sources
• Mapping Tool: A cross-reference database (e.g., myfood24's mapped dataset) linking branded products to generic food codes [34]
• Software: Data scraping tools (if needed), database management software

Procedure:

• Product Identification: Identify target product categories and brands. Use barcode scanning (if primary data collection) or database keyword searches.
• Data Extraction: For each product, extract all available label data: energy, macronutrients, sugars, sodium, fiber, and any voluntarily added micronutrients [34].
• Ingredient List Analysis: Machine-scan or manually code the ingredient list to infer fortification patterns and identify key ingredient categories [25].
• Micronutrient Imputation: Map each branded product to a comparable generic food code in a comprehensive FCDB (e.g., UK's CoFID via myfood24's method) to infer missing micronutrient values [34].
• Account for Reformulation: Check the database's "last updated" date and be aware that nutrient profiles, especially for sodium and sugars, may change frequently due to reformulation [34].
• Create Research Dataset: Compile the branded data and imputed micronutrient values into a structured dataset, flagging which values are measured versus imputed.

Protocol 3: Validating Database Quality for Nutrient Profiling Research

Purpose: To evaluate the suitability and quality of a chosen nutrient database for a specific nutrient profiling application.

Materials:

• Target nutrient database
• Reference database or scientific literature with validated analytical data
• Statistical analysis software

Procedure:

• Assess Nutrient Coverage: Calculate the percentage of missing values for each nutrient required by the profiling model (e.g., iron, zinc, vitamin A, added sugars) across the target food list.
• Check Temporal Relevance: Confirm the update cycle of the database and note if it aligns with the time frame of the research study, especially for rapidly reformulating product categories [34].
• Perform Plausibility Checks: For a random sample of foods, compare nutrient values against a high-quality reference database (e.g., USDA Foundation Foods with analytical data). Flag values with deviations greater than 20% for investigation.
• Test Model Sensitivity: Run the nutrient profiling model with the sourced data. Conduct a sensitivity analysis by replacing a subset of values with those from an alternative data source and quantify the impact on model outputs (e.g., percentage of foods changing category) [5].

Visualization of Research Workflows

The following diagrams map the logical pathways for navigating nutrient databases and integrating data for profiling models.

Diagram 1: Nutrient Database Selection Workflow

Diagram 2: Branded Food Data Augmentation Pathway

Table 3: Key Research Reagent Solutions for Nutrient Data Sourcing

Resource/Solution	Function	Research Application
USDA FoodData Central API [35]	Programmatic access to multiple USDA dataset types.	Automated data extraction for large-scale food lists and integration into analytical pipelines.
FNDDS Search Tool 2021-2023 [36]	Provides nutrient amounts for commonly consumed portions.	Standardizing food intake data from dietary recalls for national surveillance and cohort studies.
INFOODS/FAO Databases [25]	International food composition data compilation.	Cross-border studies, research in low/middle-income countries, and global nutrition comparisons.
Integrated Databases (e.g., myfood24) [34]	Pre-mapped branded products to generic food composition data.	Efficiently obtaining estimated micronutrient profiles for branded product analysis without manual mapping.
Bioanalytical Techniques (e.g., GC-MS, HPLC) [38]	Laboratory analysis of specific nutrient components.	Generating primary composition data for novel traditional varieties or validating existing database entries.

The strategic selection and application of nutrient databases is a foundational step in developing robust nutrient profiling models for comparing traditional and modern foods. National FCDBs provide comprehensive micronutrient data essential for assessing dietary adequacy, while branded databases offer real-time insights into a rapidly evolving food supply but often lack micronutrient depth. Researchers can navigate this dichotomy by employing the detailed protocols provided—particularly the data augmentation pathway for branded foods. The evolving landscape of dynamic nutrient profiling, which integrates real-time data and artificial intelligence, points toward a future where these data streams may be more seamlessly integrated, offering unprecedented precision in public health nutrition research [17].

Nutrient profiling (NP) models are quantitative algorithms that evaluate the healthfulness of foods and beverages based on their nutritional composition. These models serve as critical tools for underpinning public health policies, including front-of-pack labeling, regulation of food marketing to children, and product reformulation [2]. A fundamental distinction in NP model design lies in their approach to food categorization: across-the-board models apply identical nutrient criteria to all foods, while category-specific models employ distinct criteria tailored to different food categories [39]. This classification system represents a core methodological choice that directly influences a model's performance, applicability, and alignment with specific public health goals.

The selection between these approaches carries significant implications for nutritional research, particularly when investigating traditional versus modern food varieties. Category-specific models acknowledge the distinct nutritional roles that different foods play in the diet, potentially offering a more nuanced framework for comparing heritage and contemporary formulations within the same food group. Conversely, across-the-board models provide a unified standard that can highlight fundamental nutritional differences across diverse food types, which may be valuable when assessing overall dietary patterns or making cross-category comparisons [39] [40]. Understanding the theoretical foundations, practical applications, and empirical evidence supporting each approach is therefore essential for researchers designing studies on nutritional quality assessment.

Core Conceptual Frameworks and Definitions

Across-the-Board Nutrient Profiling Models

Across-the-board nutrient profiling models utilize a single set of nutrient criteria applied uniformly to all foods and beverages, regardless of their product category or dietary role [39]. This "one-size-fits-all" approach operates on the principle that fundamental nutritional standards for nutrients to limit (e.g., saturated fat, sodium, added sugars) and nutrients to encourage (e.g., fiber, protein, vitamins) should be consistent across the entire food supply. The theoretical foundation for this approach rests on the concept of food displacement in dietary patterns, whereby improving diet quality involves consuming more of universally healthy foods and less of universally less-healthy foods [39].

The primary advantage of across-the-board models lies in their simplicity and strong theoretical alignment with dietary guidance principles that advocate for population-wide shifts in food consumption patterns. These models provide clear, consistent signals to consumers and manufacturers about the nutritional quality of products without requiring contextual understanding of food categories. However, this approach faces the significant challenge of fairly evaluating foods with inherently different compositional profiles, potentially penalizing nutrient-dense foods that naturally contain higher levels of certain nutrients to limit (e.g., nuts with naturally occurring fats, dairy with natural sugars) [40]. This limitation has led to the development and adoption of category-specific approaches in many applied settings.

Category-Specific Nutrient Profiling Models

Category-specific nutrient profiling models employ distinct nutrient criteria tailored to different categories of foods and beverages [39]. This approach recognizes that foods play different roles in the diet and have varying inherent nutritional compositions based on their physical, chemical, and functional properties. The theoretical foundation for category-specific models aligns with the concept of food substitution within categories, where improving diet quality involves selecting healthier alternatives within the same food category [39].

Category-specific models address a key limitation of across-the-board systems by accounting for the inherent nutritional differences between food categories, thus enabling more contextually appropriate evaluations. For instance, criteria for beverages reasonably differ from those for solid foods, and standards for dairy products may justifiably vary from those for cereals. Evidence supporting this approach comes from dietary pattern research showing that individuals with healthier diets not only consume different proportions of food categories but also select healthier versions within categories [39] [41]. A study of British adults found that those with healthier diets consumed "healthier versions" of foods within meat, dairy, and cereal categories, not just different category proportions [39]. However, category-specific models introduce greater complexity and require careful definition of category boundaries, which can sometimes lead to gaming of the system through strategic product categorization.

Hybrid and Progressive Models

Emerging approaches in nutrient profiling include hybrid and progressive models that incorporate elements of both across-the-board and category-specific systems. Progressive models such as the PepsiCo Nutrition Criteria (PNC) establish category-specific frameworks with multiple achievement levels (e.g., Class IV to Class I), allowing for stepwise improvements in nutritional quality within product categories [42]. These systems recognize technical and consumer acceptance constraints while providing a roadmap for incremental reformulation.

The Food Compass 2.0 model represents another innovative approach, incorporating multiple holistic domains including nutrient ratios, food ingredients, and processing characteristics while applying uniform scoring criteria across diverse food categories [5]. This system aims to balance universal standards with sufficient nuance to accommodate category differences through its multi-domain scoring system rather than through separate category thresholds.

Table 1: Comparison of Nutrient Profiling Model Typologies

Characteristic	Across-the-Board Models	Category-Specific Models	Progressive/Hybrid Models
Theoretical Basis	Food displacement	Food substitution	Incremental improvement
Nutrient Criteria	Identical for all foods	Tailored to food categories	Category-specific with multiple tiers
Complexity	Low	High	Moderate to High
Implementation	Straightforward	Requires clear category definitions	Requires tiered criteria development
Best Application	General health guidance, simple labeling	Regulatory policies, targeted interventions	Industry reformulation, product development
Limitations	May penalize inherently nutrient-dense foods	Potential for category manipulation	Complex to develop and maintain

Decision Framework for Model Selection

Selecting between across-the-board and category-specific nutrient profiling models requires careful consideration of the research objectives, policy goals, and practical constraints. The following decision framework provides guidance for researchers and policymakers in choosing the most appropriate approach for their specific context.

Alignment with Research and Policy Objectives

The primary consideration in model selection should be alignment with the fundamental goals of the research or intervention. Evidence suggests that category-specific models with a limited number of categories are most appropriate when the aim is to promote achievable healthy dietary patterns through product reformulation and within-category substitutions [39] [41]. This approach effectively addresses the finding that healthier diets include both different proportions of food categories and healthier versions within categories.

Conversely, across-the-board models may be preferable when the objective is to provide simple, general guidance on overall food healthfulness or when implementing broad public health initiatives targeting major dietary pattern shifts. These models perform well in contexts where simplicity and consistency are prioritized over nuanced category distinctions. Research indicates that models using a large number of highly specific categories become unhelpful for promoting a healthy diet, suggesting that excessive categorization should be avoided [39].

Consideration of Nutritional Context and Population Needs

The appropriate model choice may vary depending on the nutritional context and specific population needs. In low- and middle-income countries (LMICs) facing the double burden of malnutrition, modified approaches may be necessary. For instance, "Choices" schemes that incorporate positive messages and category-specific micronutrients have been implemented in Southeast Asia and Zambia where over- and undernutrition coexist [8]. These systems encourage consumption of category-specific vitamins and minerals while advocating limiting certain nutrients.

In contexts where overnutrition is the primary concern, such as many Latin American countries, warning label systems that strongly discourage consumption of energy-dense products have been effectively implemented [8]. These systems typically use across-the-board criteria for key nutrients of concern but may incorporate category-specific exemption thresholds to account for inherently nutrient-dense foods.

Table 2: Model Selection Guidance Based on Application Context

Application Context	Recommended Approach	Rationale	Examples
Front-of-Pack Labeling	Category-specific with limited categories	Balances simplicity with category appropriateness	Nutri-Score, Health Star Rating
Marketing Restrictions	Across-the-board or broad category-specific	Provides clear, consistent standards for regulation	UK Ofcom model, WHO Regional Office models
Product Reformulation	Progressive category-specific	Incentivizes incremental improvements within categories	PepsiCo Nutrition Criteria [42]
Traditional vs. Modern Food Research	Category-specific	Enables fair comparison within food categories	Food Compass (category-adjusted scoring) [5]
Low- and Middle-Income Countries	Context-adapted category-specific	Addresses dual nutrition burdens and local priorities	National adaptations of WHO models [8]

Experimental Protocols for Model Implementation and Validation

Protocol 1: Comparative Analysis of Traditional and Modern Food Varieties

Objective: To evaluate the nutritional differences between traditional and modern varieties within specific food categories using appropriate nutrient profiling models.

Materials and Methods:

• Food Samples: Representative traditional and modern varieties from target food categories (e.g., grains, legumes, vegetables)
• Nutritional Composition Data: Obtain through chemical analysis or reliable food composition databases
• NP Models: Select both across-the-board (e.g., Food Compass 2.0) and category-specific models (e.g., category-adjusted scores) relevant to research context [5]

Procedure:

• Categorize food samples into appropriate food categories based on botanical, culinary, and functional characteristics
• Calculate nutrient profile scores for each sample using selected across-the-board and category-specific models
• Compare scores within categories to identify nutritional differences between traditional and modern varieties
• Analyze consistency of findings across different model types
• Conduct statistical analyses to determine significance of observed differences

Data Interpretation: Focus on patterns of difference that persist across multiple modeling approaches, as these likely represent robust nutritional distinctions rather than methodological artifacts.

Protocol 2: Validation Against Health Outcomes in Population Studies

Objective: To validate nutrient profiling model performance against health outcomes in diverse populations.

Materials and Methods:

• Population Data: Dietary intake data from cohort studies or national surveys with linked health outcomes
• Food Matching: Comprehensive nutrient composition database
• Health Outcomes: Clinically relevant endpoints (e.g., BMI, blood pressure, lipid profiles, disease incidence)

Procedure:

• Calculate individual-level dietary scores by applying selected NP models to food consumption data
• Assess relationships between dietary scores and health outcomes using appropriate statistical models
• Compare predictive validity of across-the-board versus category-specific approaches
• Stratify analyses by population subgroups to assess differential performance
• Evaluate model calibration and discrimination using established metrics

Validation Metrics: Include correlation coefficients, hazard ratios per standard deviation increase in score, and area under the curve for disease classification [5].

Research Reagent Solutions for Nutrient Profiling Studies

Table 3: Essential Research Tools for Nutrient Profiling Studies

Research Reagent	Function	Application Notes
Food Composition Databases	Provide nutritional data for scoring	Ensure compatibility with local food supply; consider traditional varieties
Nutrient Profiling Algorithms	Calculate healthfulness scores	Validate open-source code implementations; document modifications
Dietary Assessment Tools	Measure food consumption patterns	Adapt for traditional food intake assessment
Statistical Analysis Software	Model relationships between scores and outcomes	R, Python, or SAS with appropriate nutritional epidemiology packages
Food Categorization Frameworks	Classify foods for category-specific models	Develop clear, operational definitions for traditional food categories

Visual Decision Framework for Model Selection

The following diagram illustrates the decision-making process for selecting between across-the-board and category-specific nutrient profiling models:

Diagram 1: Model Selection Decision Tree - This workflow provides a systematic approach for researchers to select the most appropriate nutrient profiling model type based on their specific objectives and constraints.

The choice between across-the-board and category-specific nutrient profiling models represents a fundamental methodological decision with significant implications for research outcomes and policy effectiveness. Evidence suggests that category-specific models with a limited number of categories generally provide the most appropriate framework for promoting achievable healthy diets, particularly when investigating differences between traditional and modern food varieties [39] [41]. However, the optimal approach depends critically on the specific research questions, policy goals, and practical implementation constraints.

Future developments in nutrient profiling should focus on creating flexible, validated systems that can accommodate both universal nutritional principles and category-specific considerations. The integration of traditional nutritional knowledge with digital platforms presents promising opportunities for developing more culturally responsive profiling systems that respect traditional food heritage while providing scientifically rigorous nutritional evaluation [43]. As the field advances, maintaining a balance between scientific accuracy, practical applicability, and cultural appropriateness will be essential for developing nutrient profiling models that effectively serve diverse research and public health needs.

Nutrient Profiling Models (NPMs) are scientific tools that classify foods based on their nutritional composition to support public health policies, labeling schemes, and product reformulation [8]. This case study applies NPMs within a broader research thesis investigating the nutritional landscapes of traditional meat products and modern plant-based meat analogues (PBMAs). The shift toward plant-based diets, driven by health, environmental, and ethical concerns, has spurred a rapid market expansion, projected to reach €34 billion by 2027 [44]. However, the nutritional quality of these modern alternatives varies significantly, necessitating a standardized, evidence-based assessment to inform consumers, industry, and regulators [45]. This study details the protocols for a cross-sectional comparison, utilizing the UK NPM to evaluate the healthiness of meat products and their PBMA counterparts available on the retail market.

Application Notes: Theoretical and Regulatory Framework

Role of Nutrient Profiling in Public Health

NPMs serve as the foundation for various public health initiatives. They provide a quantitative basis for front-of-pack labeling (FoPL), advertising restrictions, and taxation policies. Their application is crucial in navigating the "double burden" of malnutrition, as they can be tailored to address both over-nutrition (e.g., high saturated fat, sugar, salt) and under-nutrition (e.g., micronutrient deficiencies) [8]. In the context of PBMAs, NPMs offer an objective means to evaluate their role in sustainable dietary transitions and to determine if they constitute healthier alternatives to the traditional meat products they aim to replace [46].

The UK Nutrient Profile Model

This case study employs the UK NPM, developed by the Food Standards Agency (FSA). It is a point-based system that calculates a score from the nutrient content per 100g of a product [46]. The model classifies nutrients into two groups:

• 'A' nutrients (to limit): Energy, saturated fat, total sugar, and sodium.
• 'C' nutrients (to encourage): Fruit, vegetable, and nut (FVN) content, dietary fibre, and protein.

The final score is calculated as: Total 'C' points - Total 'A' points. A product is classified as 'healthy' if the overall score is less than 4 [46]. This model's reliability is widely accepted and has been used in numerous nutrition research methodologies.

Experimental Protocol: Cross-Sectional Market Analysis

Research Workflow

The following diagram outlines the key stages of the experimental protocol for data collection and analysis.

Detailed Methodology

Data Collection Protocol

• Study Design: Cross-sectional market survey.
• Location & Timing: Data is collected in-store and online from major supermarket chains. The collection period should be clearly defined (e.g., November 2022 - March 2023) to ensure a market snapshot [44].
• Product Identification:
  • Meat Products: Include processed red meat, poultry, and fish products (e.g., burgers, sausages, bacon, minced meat, meatballs, ready meals) [46] [44].
  • Plant-Based Alternatives: Identify products explicitly marketed as direct analogues to the selected meat products, using keywords like 'vegan', 'meat-free', and 'plant-based' in conjunction with meat-associated terms (e.g., 'vegan sausages', 'plant-based burger') [46] [44]. Exclude traditional vegetarian foods not designed to mimic meat (e.g., tofu blocks, falafel).
• Sample Size: Aim for a paired design. For instance, collect 62 meat products and 62 corresponding PBMA counterparts [46].
• Data Extraction: Systematically record the following from product packaging:
  • Brand and product name.
  • Back-of-Pack (BoP) nutrition information: Energy (kcal), fat, saturated fat, carbohydrates, sugars, protein, fibre, and salt (all per 100g).
  • Front-of-Pack (FoP) label: Traffic light colour coding for fat, saturated fat, sugars, and salt.
  • Ingredient list.
  • Nutrition and health claims.

For products lacking FoP labels, colour coding should be determined post-collection using official guidance (e.g., UK Department of Health) [46].

Data Processing and Categorization

• Categorize Products: Group meat products into sub-categories: red meat, poultry, and fish. Create corresponding sub-categories for PBMAs.
• Input Data: Create a database (e.g., Microsoft Excel) with all extracted nutritional information and claims.
• Estimate FVN Content: For the UK NPM, if the quantitative ingredient declaration (QUID) is unavailable, estimate the percentage of fruit, vegetable, and nut (FVN) content based on their position in the ingredient list, following established guidance [46].

Nutrient Profiling Analysis

• Tool: Use the University of Leeds' NPM Online Calculator or an equivalent, validated computational script [46].
• Procedure:
  • Input the collected data for energy, saturated fat, sugars, sodium, fibre, and protein for each product.
  • Input the estimated FVN percentage.
  • Calculate the total 'A' points and total 'C' points for each product.
  • Compute the final NPM score (Total 'C' points - Total 'A' points).
  • Classify each product as 'healthy' (score < 4) or 'less healthy' (score ≥ 4).

Statistical Analysis

• Use statistical software (e.g., R, SPSS).
• Employ descriptive statistics to summarize the nutritional composition.
• Use inferential statistics (e.g., t-tests, ANOVA) to compare mean nutrient values and NPM scores between meat products and PBMAs, and within sub-categories.
• A significance level of p < 0.05 is typically applied. For example, a chi-squared test can assess the association between product type (meat/PBMA) and healthiness classification, with a specific focus on sub-categories like red meat [46].

Key Data Outputs and Visualization

Table 1: Mean nutritional composition per 100g for meat products versus plant-based meat alternatives (PBMAs), based on UK market data [46] and European comparative analysis [44].

Nutrient	Meat Products (Mean)	PBMA (Mean)	Significance (p-value)	Notes
Energy (kcal)	Higher	Lower	Varies by market	[44]
Protein	Higher	Lower	p = 0.029	Meat is a more concentrated source [46].
Total Fat	Higher	Lower	p = 0.029
Saturated Fat	Higher	Lower	p = 0.029	A key differentiator for cardiovascular health [46] [44].
Carbohydrates	Lower	Higher	p < 0.001
Dietary Fibre	Low or zero	Significantly Higher	p < 0.001	A distinct advantage of PBMAs [46] [44].
Salt	Varies	Varies, often higher	Not always significant	Highly dependent on product formulation [44].
Cholesterol	Present	Absent	N/A	PBMAs contain no dietary cholesterol [47].

Nutrient Profiling and Claim Outcomes

Table 2: Results from NPM application and claim analysis from a UK market study (n=62 per category) [46].

Metric	Meat Products	Plant-Based Alternatives
Proportion Classified 'Healthy' (NPM Score <4)	60%	74%
Association with Healthiness (Red Meat sub-category)	Significant negative association (p=0.005)	Healthier profile in this category
Products with Nutrition Claims	15%	40%
Common Nutrition Claims	--	Protein content (34% of red meat PBMAs)
Products with Health Claims	0%	0%
FoP Red Coding (High content)	More prominent for fat (23%) and saturated fat (35%)	Less prominent for fat and saturated fat

The Scientist's Toolkit: Research Reagent Solutions

Table 3: Essential materials and tools for conducting NPM comparative research.

Item / Tool	Function / Purpose
UK NPM Online Calculator	Web-based tool for calculating and classifying products based on the UK NPM algorithm [46].
Statistical Software (R, SPSS, SAS)	To perform descriptive and inferential statistical analysis on nutritional data and NPM scores.
Data Collection Database (Excel, SQL)	To systematically record, store, and manage product information and nutritional data extracted from labels.
Department of Health FoP Guidance	Official document to assign traffic light colour coding to products that lack FoP labels [46].
QUID/FVN Estimation Protocol	A standardized method for estimating the percentage of fruit, vegetable, and nut content from ingredient lists when precise QUID is unavailable [46].
Supermarket Access (Online/Physical)	Provides the sample population of products for this market-based research.

The integration of Indigenous knowledge with modern nutritional science presents a transformative opportunity for public health and food research. Indigenous knowledge systems encompass profound understandings of local ecosystems, traditional food sources, and their health implications, developed over millennia. This document provides application notes and protocols for researchers to ethically document and digitize this knowledge, with specific emphasis on developing nutrient profiling models (NPMs) that accurately represent traditional food varieties. Such integration is crucial for addressing the dual challenges of malnutrition and biodiversity loss, while respecting the rights of Indigenous Peoples as knowledge holders. Framing this within nutrient profiling research creates a robust bridge between empirical Western science and holistic traditional knowledge systems, enabling more culturally relevant and effective public health nutrition strategies.

International Policy and Ethical Frameworks

Key International Instruments

Research involving Indigenous knowledge must be grounded in recognized ethical frameworks that prioritize rights, equity, and benefit-sharing. The following table summarizes the core international instruments relevant to this work.

Table 1: Key International Frameworks Governing Indigenous Knowledge

Framework/Instrument	Governing Body	Core Principles	Relevance to Nutrient Profiling
Convention on Biological Diversity (CBD) Article 8(j) [48]	UN CBD	Respect, preserve, and maintain knowledge; promote wider application with approval and involvement of holders; encourage equitable benefit-sharing [48].	Mandates ethical use of traditional knowledge about food biodiversity and its health benefits.
WIPO Treaty on Intellectual Property, Genetic Resources and Associated Traditional Knowledge (2024) [49]	World Intellectual Property Organization (WIPO)	First WIPO treaty with specific provisions for Indigenous Peoples and local communities; addresses IP interface with genetic resources and traditional knowledge [49].	Provides guidance on IP protection for traditional knowledge related to food and medicinal plants.
UNESCO Indigenous Knowledge, Ancestral Places (2025) [50]	UNESCO	Highlights Indigenous knowledge systems as dynamic, evolving skills for addressing modern challenges, including sustainable stewardship [50].	Recognizes traditional knowledge as a valid system for understanding food systems and environmental health.

Core Ethical Protocols

Prior to any documentation activity, researchers must:

• Obtain Free, Prior, and Informed Consent (FPIC): Engage with appropriate community governance structures to secure collective consent, which is ongoing and can be withdrawn at any stage.
• Establish Mutually Agreed Terms (MAT): Co-develop written agreements covering project goals, data ownership, access, use, confidentiality, and benefit-sharing.
• Respect Cultural Protocols: Adhere to community-specific rules regarding the sharing of sacred or restricted knowledge. Some knowledge may not be suitable for digitization or wide dissemination.
• Ensure Equitable Benefit-Sharing: Benefits, which may be monetary (e.g., royalties) or non-monetary (e.g., capacity building, shared research data, community recognition), must be fairly negotiated.

Phase 1: Community Engagement & Knowledge Documentation Protocols

This phase focuses on building trust and ethically recording knowledge in partnership with communities.

Community Engagement Workflow

The following diagram illustrates the critical, iterative pathway for ethical community engagement.

Documentation Protocol: Ethnobotanical Surveys for Nutrient Profiling

Aim: To systematically document traditional food plants, their culinary uses, and perceived health benefits within a cultural context, providing qualitative data for NPM development.

Materials:

• Digital voice recorder, GPS device, calibrated digital scales, high-resolution camera, field notebooks, standardized data sheets.
• Prior Informed Consent (PIC) forms, material transfer agreement (MTA) templates, benefit-sharing agreement.

Methodology:

• Semi-Structured Interviews:
  • Selection: Engage knowledge holders (e.g., elders, healers, farmers) identified through community leadership, using purposive snowball sampling.
  • Questionnaire: Use a flexible guide covering:
    • Local name(s) of food.
    • Phenological description (if plant/animal is not present).
    • Seasonal availability.
    • Traditional harvesting/cultivation practices.
    • Methods of preparation, processing, and preservation.
    • Perceived health benefits and associated cultural beliefs.
    • Use for specific demographic groups (e.g., children, pregnant women, elderly).
  • Context: Conduct interviews in the preferred language of the participant, with the assistance of a trained community translator.

• Focus Group Discussions (FGDs):

  • Convene groups of 6-10 individuals to discuss community-level knowledge, such as seasonal food calendars, festive foods, and consensus on food properties.
  • Use participatory rural appraisal (PRA) tools like seasonal calendars and pairwise ranking to generate comparative data on food preferences and health perceptions.

• Field Collection & Voucher Specimens:

  • With permission, collect plant specimens in triplicate with a knowledge holder.
  • Record GPS coordinates, habitat data, and photograph the specimen in situ.
  • Prepare voucher specimens according to standard herbarium protocols. One set is deposited in a recognized national herbarium, and a duplicate set is returned to the community.

• Data Management:

  • All recordings are transcribed and, where necessary, translated.
  • Data is entered into a standardized database (e.g., using Excel or a custom SQL database) linked to voucher specimen numbers and media files.

Phase 2: Digitization & Data Integration Protocols

This phase converts documented knowledge into structured, analyzable data while preserving context and upholding ethical data governance.

Digitization and Modeling Workflow

The following diagram outlines the technical process for creating a integrated data model for traditional food knowledge.

Protocol: Creating a Traditional Knowledge Digital Repository

Aim: To create a secure, searchable database for documented traditional knowledge that respects cultural protocols and enables integration with biochemical data.

Materials: Server infrastructure (preferably with community-level access control), database management software (e.g., PostgreSQL, MySQL), metadata schemas (e.g., Dublin Core Extended for TK), WIPO TK Toolkit resources [49].

Methodology:

• Data Curation and Annotation:
  • Transcribed and translated data is tagged using a controlled vocabulary (e.g., on health benefits: "energy," "digestion," "lactation").
  • Each data entry is linked to its source (interview ID, knowledge holder) and associated voucher specimen.

• Database Schema Design:

  • Design tables to store information on Taxa, Uses, Recipes, Knowledge Holders, and Collections.
  • Implement user-based permissions to control access to sensitive or confidential fields based on community-defined rules.
  • Use standardized metadata fields to track provenance, access restrictions, and licensing conditions.

• Integration with Nutritional Data:

  • The Taxa table in the TK database must include a field to link to a separate Nutritional_Composition table.
  • The Nutritional_Composition table stores quantitative data on proximate analysis, vitamins, minerals, and bioactive compounds derived from laboratory analysis.

Application in Nutrient Profiling Model (NPM) Research

The integrated data model enables the development of nuanced NPMs that reflect the value of traditional foods.

Protocol: Adapting NPMs for Traditional Food Varieties

Aim: To adapt or create a nutrient profiling model that incorporates the unique attributes of traditional foods, moving beyond standard negative nutrient limits (sugar, fat, salt) to include positive aspects like food processing and cultural context.

Background: Modern NPMs like Food Compass 2.0 are evolving to include attributes like food processing and specific ingredients [5]. Region-specific models like the ANPS-Meal for Japan focus on public health priorities, using components like protein, vegetables, saturated fat, and sodium [51]. Similarly, "Choices" schemes in Southeast Asia and Zambia encourage category-specific vitamins and minerals where over- and undernutrition coexist [8]. These principles can be extended to value traditional knowledge.

Methodology:

• Component Selection:
  • Positive Components: Identify nutrients of interest from biochemical analysis (e.g., specific bioactive compounds, dietary fiber, micronutrients prevalent in traditional varieties).
  • Negative Components: Standard nutrients to limit (e.g., sodium, saturated fat). Use data from laboratory analysis.
  • Processing & Ingredient Attributes: Incorporate a positive score for traditional, minimal processing methods (e.g., fermentation, sun-drying) as recognized in Food Compass 2.0 [5].
  • Cultural Use Score: Develop a qualitative score based on documented ethnobotanical data for foods of high cultural significance or specific health applications.

• Scoring Algorithm:

  • Adapt an existing model (e.g., a simplified Food Compass or ANPS algorithm) or create a new one.
  • Example Framework: Total Score = (Positive Nutrients Score) - (Negative Nutrients Score) + (Traditional Processing Bonus) + (Cultural Use Score)
  • Normalize scores to a 0-100 scale for comparability.

• Validation:

  • Convergent Validity: Correlate the new model's scores with established indices like the Healthy Eating Index-2015 (HEI-2015) or NRF9.3, as done for ANPS-Meal (r=0.59 with mHEI-2015) [51].
  • Face Validity: Present the ranked foods and scores to community knowledge holders for verification and discussion, ensuring the model's outputs align with cultural understanding.

Comparative Analysis Framework

Table 2: Framework for Comparing Traditional and Modern Varieties Using an Adapted NPM

Analysis Dimension	Data Input (Traditional Variety)	Data Input (Modern Variety)	Method of Comparison
Macronutrient Profile	Proximate analysis from lab data.	Secondary data from food composition tables.	T-test/Wilcoxon rank-sum test; difference in means.
Micronutrient Density	Quantified vitamins & minerals per 100g or 100kcal.	Secondary data from food composition tables.	Calculate NPM sub-score for positive nutrients; compare.
Bioactive Compounds	Targeted analysis for specific phytochemicals (e.g., flavonoids, polyphenols).	Literature data, if available.	Qualitative and quantitative profiling; incorporate into NPM if a positive component is defined.
Processing Score	Score based on traditional methods (e.g., +5 for fermentation).	Score based on industrial methods (e.g., -5 for ultra-processing).	Compare attribute scores within the adapted NPM.
Overall Healthfulness	Total score from the adapted NPM.	Total score from the adapted NPM.	Statistical comparison of total NPM scores between groups.

The Scientist's Toolkit: Essential Research Reagents & Materials

Table 3: Key Reagents and Materials for Documenting and Analyzing Traditional Foods

Item/Category	Specification/Example	Primary Function in Research
Voucher Specimen Collection	Plant press, silica gel, herbarium sheets, GPS logger, taxonomic keys.	Provides a permanent, verifiable link between a local name, a physical specimen, and its scientific identification.
Bioactive Compound Analysis	HPLC-MS/MS, Standard reference materials (e.g., polyphenol standards).	Identifies and quantifies specific phytochemicals that may underpin traditional health claims.
Proximate Analysis	Soxhlet apparatus (fat), Kjeldahl apparatus (protein), muffle furnace (ash).	Determines the basic macronutrient and mineral composition of food samples.
Metadata Standards	WIPO TK Toolkit [49], Dublin Core Extended.	Ensures traditional knowledge is documented with proper provenance, rights, and access conditions.
Ethical Agreement Templates	Free, Prior, and Informed Consent (FPIC) form, Material Transfer Agreement (MTA), Benefit-Sharing Agreement.	Provides the legal and ethical foundation for research, protecting both the community and the researchers.
Nutrient Profiling Algorithm	Adapted Food Compass 2.0 [5] or ANPS-Meal [51] algorithm.	The core analytical model for quantitatively assessing and comparing the healthfulness of foods.

Overcoming Profiling Pitfalls: Technical, Cultural, and Data-Limitation Challenges

The digitization of dietary guidance represents a significant public health opportunity; however, current digital nutrition platforms predominantly utilize standardized nutrient profiling models developed for Western contexts with modern food supplies. This creates a critical gap in accurately representing traditional food systems, which are often rich in biodiversity, cultural significance, and context-specific nutritional benefits. This application note details protocols for identifying these representation gaps and proposes methodological frameworks for integrating traditional food knowledge into digital nutrition platforms, specifically within research contexts examining nutrient profiling models for traditional versus modern varieties.

Nutrient profiling (NP) models provide the algorithmic foundation for most digital nutrition platforms, initially created to prevent obesity in high-income countries by penalizing foods containing excessive calories, fat, sugar, and salt [25]. These energy-driven NP models have significant limitations when applied to traditional food systems in low- and middle-income countries where undernutrition and micronutrient deficiencies remain pressing concerns [25]. Current digital nutrition platforms often fail to reflect dietary diversity, relying instead on standardized models with limited cultural sensitivity [43]. This underrepresentation has profound implications for nutritional science, public health interventions, and the preservation of traditional knowledge systems, particularly as digital platforms become primary sources of dietary guidance for researchers, healthcare professionals, and consumers.

Critical Analysis of Existing Nutrient Profiling Models

Limitations for Traditional Food Assessment

Table 1: Limitations of Current NP Models for Traditional Foods

NP Model Characteristic	Modern Food Application	Traditional Food Limitation	Impact on Accuracy
Nutrients Evaluated	Primarily nutrients to limit (sodium, saturated fat, total sugars) [52]	Overlooks beneficial nutrients (iron, zinc, calcium, vitamins) often deficient in traditional diet contexts [25]	Underestimates nutritional value of traditional foods addressing specific deficiencies
Validation Approach	Content or face validity testing conducted for only 42% of models [52]	Lack of validation against traditional dietary patterns and health outcomes	Limited evidence base for application in diverse populations
Food Categorization	Category-specific models favor processed food comparisons [25]	Inappropriate grouping of traditional foods with dissimilar modern products	Misclassification of traditional foods within nutritional hierarchies
Fortification Handling	Variable recognition of fortified foods [25]	Failure to account for traditional bioavailability enhancement techniques (e.g., fermentation)	Incomplete nutritional assessment of traditional preparation methods
Protein Quality	Generally excluded from high-income country models [25]	Overlooks protein quality issues in staple traditional foods (e.g., cassava) [25]	Significant gap in assessing traditional food systems with protein limitations

Digital Platform Implementation Deficits

Current digital nutrition platforms demonstrate significant technical and cultural barriers to traditional food integration. Evaluation of five leading digital nutrition platforms (MyFitnessPal, Cronometer, Noom, Lifesum, and Nutritics) reveals limited representation of diverse traditional foods, minimal cultural adaptability, and absence of contextual information about traditional dietary patterns [43]. The underlying data architectures of these platforms typically lack flexibility to incorporate culturally-specific food classifications, preparation methods, and holistic nutritional concepts embedded in traditional food wisdom [43].

Experimental Protocols for Gap Identification and Resolution

Protocol 1: Documentation of Traditional Food Systems

Objective: Systematically document traditional food items, their nutrient profiles, and cultural contextual data for digital integration.

Materials and Reagents:

• Local nutrient composition databases (e.g., INFOODS, SMILING database) [25]
• Food sampling kits (containers, labels, temperature control)
• Digital data collection tools (structured interview protocols, photographic equipment)
• Cultural domain analysis frameworks (PEN-3 Cultural Model) [43]

Methodology:

• Food Identification and Collection: Identify key traditional foods through community engagement with local knowledge holders. Document seasonal availability, varieties, and procurement sources.
• Nutrient Analysis: Conduct laboratory analysis of traditional food samples for comprehensive nutrient profiles, prioritizing micronutrients of public health concern in specific regions (iron, zinc, calcium, vitamin A, B vitamins) [25].
• Contextual Data Documentation: Record preparation methods, culinary uses, seasonal consumption patterns, and cultural significance through structured interviews and observation.
• Data Standardization: Format documented information according to FAO/INFOODS standards for food composition data [25].

Validation: Cross-verify nutrient data with existing scientific literature and validate cultural contextual information through community member feedback.

Protocol 2: Cultural Adaptation of Nutrient Profiling Algorithms

Objective: Modify existing NP algorithms to appropriately evaluate traditional foods while maintaining scientific rigor.

Materials and Reagents:

• Existing NP models (e.g., Nutri-Score, Choices International) [25]
• Traditional food nutrient database (from Protocol 1)
• Statistical software (R, Python with pandas, scikit-learn)
• Cultural adaptation frameworks (Bernal's model) [43]

Methodology:

• Algorithm Audit: Analyze current NP model components to identify cultural biases against traditional foods.
• Beneficial Nutrient Integration: Incorporate weighting for regionally-relevant micronutrients and high-quality protein where these represent documented public health concerns [25].
• Contextual Factors: Develop modifiers for traditional preparation methods that enhance nutrient bioavailability (e.g., fermentation, soaking).
• Validation Testing: Compare original versus adapted algorithm outputs against health outcomes in populations consuming traditional diets.
• Stakeholder Review: Submit adapted algorithms for review by both nutrition scientists and cultural knowledge holders.

Validation: Conduct content validity testing with experts in both nutritional science and traditional food systems, and assess face validity with consumer groups from traditional backgrounds.

Protocol 3: Digital Platform Integration Framework

Objective: Implement a structured approach for integrating documented traditional food data into digital nutrition platforms.

Materials and Reagents:

• Documented traditional food database (from Protocol 1)
• Adapted NP algorithms (from Protocol 2)
• Participatory design frameworks (Knowledge-to-Action Framework) [43]
• Digital platform development tools

Methodology:

• Data Architecture Modification: Expand food classification systems to accommodate traditional food categories and relationships.
• Algorithm Implementation: Integrate adapted NP models into platform backend with appropriate cultural modifiers.
• Interface Adaptation: Develop culturally-informed user interfaces that reflect traditional food knowledge organization and concepts.
• User Testing: Conduct iterative testing with diverse user groups to ensure cultural appropriateness and usability.
• Implementation Scaling: Develop protocols for regional adaptation of the platform to different traditional food systems.

Validation: Measure platform accuracy in representing traditional foods, user satisfaction across cultural groups, and adoption rates in target communities.

Visualization of Research Workflows

Traditional Food Digital Integration Pathway

Nutrient Profiling Model Adaptation Logic

Research Reagent Solutions for Traditional Food Analysis

Table 2: Essential Research Materials for Traditional Food Nutrient Profiling

Reagent/Resource	Function	Application Specifics	Quality Considerations
INFOODS Compendium	Standardized food composition data	Provides regional food data templates; reference for traditional food nutrients [25]	Regular updates required; regional validation recommended
SMILING Database	Specialized Asian food composition data	Contains traditional food data from Cambodia, Indonesia, Laos, Thailand, Vietnam [25]	Cultural specificity; limited geographic scope
USDA SR-28 Database	Comprehensive nutrient reference	Backup reference for universal nutrients; baseline comparison [25]	Cultural limitations; potential mismatches for traditional varieties
PEN-3 Cultural Model	Cultural context assessment framework	Guides documentation of cultural dimensions of traditional foods [43]	Requires cultural competency for proper application
Knowledge-to-Action Framework	Implementation science structure	Supports translation of traditional knowledge to digital formats [43]	Flexible adaptation to specific cultural contexts needed
ColorBrewer Palettes	Data visualization color schemes	Ensures accessible representation of traditional food data [53]	Colorblind-safe selections; cultural color associations

Data Presentation Standards for Traditional Food Research

Quantitative Gap Analysis Framework

Table 3: Metrics for Assessing Traditional Food Representation Gaps

Assessment Dimension	Metric	Measurement Method	Interpretation Guidelines
Database Completeness	Percentage of core traditional foods with complete nutrient profiles	Audit of existing nutrition databases against ethnographic food lists	<50% = Critical gap; 50-80% = Significant gap; >80% = Moderate gap
Algorithm Appropriateness	Number of traditional foods misclassified by current NP models	Comparison of traditional food scores against nutrition expert ratings	>30% misclassification = Critical limitation
Cultural Context Inclusion	Presence/absence of key contextual factors in digital records	Checklist assessment: seasonal use, preparation methods, cultural significance	Missing >40% factors = Serious contextual deficit
Platform Usability	User task success rate for traditional food entries	Controlled usability testing with diverse cultural groups	<70% success rate = Major usability barrier
Health Outcome Alignment	Correlation between NP scores and actual health impacts in traditional diet populations	Longitudinal studies of traditional food consumers	r<0.3 = Weak predictive validity

The underrepresentation of traditional foods in digital platforms constitutes a significant scientific and equity gap that demands methodical addressing through the protocols outlined herein. Researchers should prioritize collaborative partnerships with traditional knowledge holders, recognizing that effective digital integration requires both scientific rigor and cultural respect. The adaptation of nutrient profiling models must balance the need for global standardization with appropriate contextualization for traditional food systems. Future research directions should include longitudinal validation of adapted NP models against health outcomes in populations consuming traditional diets, development of standardized protocols for cross-cultural food documentation, and establishment of international databases for traditional food composition with appropriate cultural metadata.

Nutrient Profiling Models (NPMs) face a significant challenge in accurately characterizing both ultra-processed foods (UPFs) and nutritionally-rich traditional foods, creating a critical "processing paradox" in modern nutritional science. This application note analyzes the structural limitations of current NPMs through quantitative analysis of profiling results across diverse food categories. We present standardized experimental protocols for evaluating food processing dimensions and nutrient density, alongside mechanistic insights into how processing affects nutritional quality. Our findings reveal that contemporary NPMs frequently misclassify both UPFs with favorable nutrient profiles and traditional foods with complex nutritional matrices, highlighting the urgent need for integrated assessment frameworks that account for processing methods, food matrix effects, and cultural context in nutritional evaluation.

The fundamental challenge in nutrient profiling lies in reconciling two competing dimensions: nutrient composition versus processing characteristics. Traditional NPMs primarily assess foods based on their content of specific nutrients to encourage or limit, often overlooking how industrial processing alters food matrices, bioavailability, and physiological impacts [54] [55]. This limitation becomes particularly evident when evaluating both ultra-processed foods that meet conventional nutrient criteria and traditional whole foods with complex nutritional profiles that may not align with standardized metrics.

The NOVA classification system represents a significant advancement by categorizing foods based on processing extent and purpose rather than nutrient content alone [55] [56]. However, its qualitative nature and occasional inconsistencies have sparked scientific debate. As Hall's research demonstrates, even when macronutrients are matched, UPFs and minimally processed foods have different effects on energy intake and body weight, suggesting that processing itself introduces variables beyond nutrient composition [55] [56]. Simultaneously, traditional food systems—such as Japan's washoku, Argentina's indigenous foods, and Italy's Mediterranean diet—embody holistic nutritional wisdom that transcends simple nutrient metrics, incorporating aspects of food synergy, cultural preparation methods, and bioavailability [43].

This application note addresses the "processing paradox" by providing researchers with standardized methodologies to bridge this conceptual and methodological gap, enabling more accurate characterization of both modern and traditional foods within a unified analytical framework.

Quantitative Analysis: Comparative Performance of Profiling Systems

Table 1: Food Compass 2.0 Scoring Reveals Processing Paradox in Major Food Categories

Food Category	Original Food Compass Score (Mean ± SD)	Food Compass 2.0 Score (Mean ± SD)	Score Change	Key Drivers of Change
Cold Cereals	51 ± 21	41 ± 20	-10	Processing characteristics, added sugar, artificial additives
Seafood	72 ± 14	81 ± 14	+9	Minimally processed nature, beneficial nutrients
Beef	33 ± 6	44 ± 6	+11	Revised dairy fat assessment, processing considerations
Plant-Based Dairy Alternatives	54 ± 21	43 ± 20	-11	Additives, processing methods, nutrient quality
Eggs	46 ± 13	54 ± 13	+8	Minimally processed nature, matrix effects
Fruit & Vegetable Juices	72 ± 15	66 ± 14	-6	Added sugar as ingredient, fiber loss during processing

Data derived from Food Compass 2.0 validation study on 9,273 food items [5]

Table 2: Child-Targeted Food Analysis Reveals Widespread Nutritional Inadequacy

Assessment Method	Percentage Non-Compliant/Inadequate	Key Findings
WHO NPM-2023 Criteria	93.2%	Majority should not be marketed to children
Nutri-Score (D/E Groups)	70%	Low nutritional quality prevalent
NOVA (Ultra-Processed)	92.7%	Strong correlation with poor nutritional quality
UK FSA/Ofcom NPM	89.5%	High in fats, sugars, or salt

Data from analysis of 775 child-targeted packaged foods in Türkiye [57]

The quantitative evidence demonstrates significant discrepancies in how different profiling systems characterize food healthfulness. Food Compass 2.0's substantial revisions—particularly downward adjustments for processed plant-based alternatives and upward adjustments for minimally processed animal foods—highlight evolving understanding of processing impacts [5]. The consistent correlation between UPF classification and poor nutritional quality in child-targeted foods underscores the public health implications of these classification challenges [57].

Experimental Protocols: Methodologies for Comprehensive Food Characterization

Protocol 1: Integrated Formulation & Processing Impact Assessment

Purpose: To quantitatively dissect the separate contributions of formulation and processing to a food's nutritional value and health implications [54].

Materials:

• IUFoST Formulation & Processing Classification (IF&PC) framework
• Nutritional composition data (pre- and post-processing)
• Nutrient Rich Food (NRF) index calculator
• Processing intensity metrics

Procedure:

• Formulation Quantification (F):
  • Calculate baseline NRF index using pre-processing ingredient composition
  • Document all ingredients using systematic selection criteria
  • Quantify nutrients to encourage (protein, fiber, vitamins, minerals) and limit (saturated fat, sodium, added sugars)

• Processing Impact Assessment (P):

  • Process the formulated food using standard industrial methods
  • Analyze nutritional composition post-processing
  • Calculate ΔNRF as the difference in NRF index between pre- and post-processing states
  • Classify processing methods using standardized descriptors (mechanical, thermal, chemical, biological)

• Integrated Classification:

  • Compute FPFI_N = (F × w_F) + (ΔNRF × w_P)
  • Apply appropriate weighting factors (w_F, w_P) based on food category
  • Implement color-coded visualization scheme (green: favorable, red: unfavorable)

Validation: Compare IF&PC classification against in vitro digestibility studies and clinical postprandial responses for method validation [54].

Protocol 2: Cultural Food System Documentation and Integration

Purpose: To systematically document and integrate traditional food knowledge into digital nutrition platforms while preserving cultural context and nutritional complexity [43].

Materials:

• Cultural adaptation framework (Bernal's model)
• PEN-3 Cultural Model assessment tools
• Food composition databases (USDA, TURKOMP, traditional food databases)
• Digital documentation equipment

Procedure:

• Community Engagement Phase:
  • Identify and collaborate with cultural knowledge-keepers and community elders
  • Conduct structured interviews on traditional food preparation methods
  • Document seasonal variations and ceremonial food uses

• Nutritional Documentation:

  • Collect samples of traditional foods in various preparation states
  • Conduct laboratory analysis of macro- and micronutrient composition
  • Document food matrix characteristics and preparation techniques
  • Assess bioactive compound preservation through different processing methods

• Digital Integration:

  • Develop flexible data architecture accommodating cultural context
  • Implement culturally-informed algorithms for nutritional assessment
  • Create metadata structures for preparation methods and food combinations
  • Validate integration through community feedback and nutritional accuracy testing

Application: This protocol successfully enabled the crediting of traditional Indigenous foods in USDA Child Nutrition Programs by establishing nutritional equivalencies with conventional foods [58].

Signaling Pathways and Mechanistic Insights: Biological Interfaces of Food Processing

Mechanistic Pathways Linking UPF Consumption to Health Outcomes

The diagram illustrates the multifactorial pathways through which ultra-processed foods impact health, extending beyond their nutrient profile to include food matrix alterations, additive effects, and processing-induced contaminants [55]. These mechanisms help explain why UPFs exert physiological effects distinct from their nutrient composition alone, including:

• Gut microbiome dysbiosis: Industrial additives and processing methods alter microbial composition and function [55]
• Chronic inflammation: Modified food matrices and processing contaminants trigger inflammatory pathways [55]
• Hormonal imbalances: Rapid nutrient delivery and matrix disruption affect satiety signaling and metabolic regulation [56]

Experimental Workflow: Integrated Food Characterization

Comprehensive Food Characterization Workflow

The integrated workflow combines conventional nutrient analysis with processing assessment and cultural context documentation to address limitations of single-dimension profiling systems. This approach enables researchers to:

• Quantitatively separate formulation and processing effects [54]
• Incorporate traditional food knowledge and preparation methods [43]
• Generate multidimensional healthfulness scores validated through clinical studies [17]

Research Reagent Solutions: Essential Methodological Tools

Table 3: Key Research Reagents and Analytical Tools for Advanced Nutrient Profiling

Reagent/Tool	Application	Function	Validation Requirements
IUFoST IF&PC Framework	Processing Impact Quantification	Separates formulation and processing effects	Correlation with clinical responses [54]
NRF Index	Nutrient Density Assessment	Bases formulation scoring on beneficial vs. limiting nutrients	Comparison with health outcome data [54]
NOVA Classification	Processing Degree Categorization	Classifies foods into 4 processing categories	Inter-rater reliability testing [57]
Cultural Adaptation Models	Traditional Food Integration	Structures incorporation of indigenous knowledge	Community validation and acceptance [43]
Dynamic Profiling Algorithms	Personalized Nutrition	Adapts recommendations based on real-time data	Clinical outcomes in diverse populations [17]
Food Composition Databases	Nutritional Analysis	Provides baseline nutrient data	Analytical method standardization [58]

These research tools enable comprehensive characterization of both modern and traditional foods, addressing the processing paradox through multidimensional assessment. The IUFoST framework specifically addresses NOVA limitations by providing quantitative processing metrics, while cultural adaptation models facilitate appropriate integration of traditional food knowledge into contemporary nutritional assessment [43] [54].

Addressing the processing paradox requires moving beyond conventional nutrient-centric models toward integrated systems that account for processing methods, food matrix effects, and cultural context. The experimental protocols and analytical frameworks presented here provide researchers with standardized methodologies to dissect the complex relationships between formulation, processing, and health impacts. Future innovation should focus on dynamic profiling systems that incorporate real-time physiological data, account for individual variability in response to processed foods, and respectfully integrate traditional food wisdom [17]. Such advances will enable more accurate characterization of both traditional and modern foods, supporting evidence-based policies that promote genuine healthfulness rather than simplified nutrient metrics.

Nutrient Profiling Models (NPMs) represent a critical methodological tool in public health nutrition, enabling the classification of foods based on their nutritional composition to support health promotion and disease prevention strategies [11]. Originally developed to combat diet-related chronic diseases in Western contexts, the application of NPMs in low- and middle-income countries (LMICs) presents unique challenges due to the complex double burden of malnutrition—where undernutrition and micronutrient deficiencies coexist with rising rates of overweight, obesity, and diet-related non-communicable diseases [8] [59]. This protocol outlines a systematic approach for adapting NPMs to address these dual nutritional challenges while considering both traditional and modern food varieties.

The persistent global malnutrition crisis underscores the urgency of this work. Current data indicates that 149.2 million children under five are stunted, 45.4 million are wasted, and 38.9 million are overweight, while 2.2 billion adults experience overweight or obesity [60]. Most countries are not on track to meet global nutrition targets, necessitating innovative approaches to nutritional assessment and intervention [60]. Effective NPM adaptations must balance the need to discourage consumption of energy-dense, nutrient-poor foods while encouraging intake of nutrient-dense foods, particularly in populations experiencing multiple forms of malnutrition [8].

Current Landscape of Nutrient Profiling Models

Established NPM Frameworks

Multiple NPM frameworks have been developed and validated across different global contexts. Table 1 summarizes the key characteristics of major profiling models, highlighting their varying approaches to addressing nutritional priorities.

Table 1: Comparison of Major Nutrient Profiling Models

Model Name	Region/Country	Key Nutrients to Limit	Key Nutrients to Encourage	Primary Application	Validation Status
Ofcom	United Kingdom	Saturated fat, sodium, sugars	Protein, fiber, fruit/vegetable/nut content	Marketing restrictions to children	Substantial [11] [61]
Nutri-Score	France	Saturated fat, sodium, sugars	Protein, fiber, fruits/vegetables/legumes/nuts	Front-of-pack labeling	Substantial criterion validation [6]
Health Star Rating	Australia/New Zealand	Saturated fat, sodium, sugars	Protein, fiber, fruit/vegetable content	Front-of-pack labeling	Intermediate validation [6]
PAHO	Americas	Free sugars, saturated fat, trans fat, sodium	None specified	Policy development	Limited validation [11]
FSANZ	Australia/New Zealand	Saturated fat, sodium, sugars	Protein, fiber, fruit/vegetable content	Health claims regulation	Near perfect agreement with Ofcom [11]
HCST	Canada	Saturated fat, sodium, sugars	Protein	Surveillance	Fair agreement with Ofcom [11]

Global Implementation Patterns in LMICs

Implementation of NPMs in LMICs reveals distinct patterns based on prevailing nutritional challenges. Warning Label schemes, which strongly discourage consumption of energy-dense products, have been predominantly adopted in Latin American LMICs where overnutrition affects most of the population [8]. These models typically focus on limiting sugars, fats, and sodium.

In contrast, "Choices" schemes that incorporate positive messages have been implemented in Southeast Asia and Zambia, where over- and undernutrition frequently coexist [8] [59]. These systems not only advocate limiting certain nutrients but also encourage consumption of category-specific vitamins and minerals essential for addressing local micronutrient deficiencies.

The "Keyhole" front-of-pack labeling scheme, implemented in North Macedonia, represents an intermediate approach that limits sugars, fat, and salt while promoting fibers, fruits, vegetables, nuts, and legumes [8]. This balanced strategy aims to prevent overnutrition and diet-related chronic diseases while maintaining nutritional adequacy.

Adaptation Framework for Double Burden Contexts

Situational Analysis Protocol

Objective: To systematically assess the nutritional landscape and food environment to inform context-specific NPM adaptation.

Methodology:

• Epidemiological Assessment:
  • Extract prevalence data for undernutrition, overweight/obesity, and key micronutrient deficiencies (iron, vitamin A, iodine) from the Global Health Observatory and Global Burden of Disease study [8] [59]
  • Analyze trends in diet-related non-communicable diseases using national survey data and published literature
  • Identify population subgroups most vulnerable to different forms of malnutrition

• Food Supply Characterization:

  • Document the relative availability and consumption patterns of traditional versus modern food varieties
  • Assess the degree of nutrition transition through food expenditure and consumption surveys
  • Map traditional food sources of limiting nutrients and micronutrients commonly deficient in the population

• Policy Environment Mapping:

  • Inventory existing nutrition policies, programs, and regulatory frameworks
  • Identify key stakeholders and institutional capacities for NPM implementation
  • Assess potential barriers and facilitators to NPM adoption

Figure 1: Decision Pathway for NPM Adaptation in LMICs

Model Selection and Modification Protocol

Objective: To select and adapt an appropriate NPM framework based on the situational analysis findings.

Methodology:

• Base Model Identification:
  • Select an existing validated model (e.g., Ofcom, Nutri-Score) as a starting point [11]
  • Prioritize models with substantial criterion validation evidence where possible [6]
  • Consider regional relevance and similarity of food supply

• Contextual Modification:

  • For contexts with significant undernutrition: Incorporate encouragement factors for locally relevant micronutrients (iron, zinc, vitamin A, iodine) and protein [8] [59]
  • For contexts with predominant overnutrition: Strengthen limits on energy-dense nutrients (saturated fats, added sugars, sodium) while maintaining emphasis on dietary quality
  • For mixed contexts: Implement balanced approaches that both limit negative nutrients and encourage positive food components and micronutrients

• Validation and Testing:

  • Assess content validity through expert consultation on nutrient inclusion
  • Conduct construct/convergent validation against a reference standard [11]
  • Test classification outcomes across food categories, with particular attention to traditional foods and meals

Experimental Protocols for NPM Validation

Content Validity Assessment Protocol

Objective: To evaluate whether the adapted NPM encompasses the full range of meaning for healthfulness within the specific context.

Materials:

• Comprehensive nutrient composition database
• Food consumption pattern data
• Expert panel with knowledge of local nutrition priorities

Procedure:

• Nutrient Selection Justification:
  • Document public health rationale for each included nutrient based on local burden of disease data
  • Ensure the model addresses both nutrients to limit (based on overnutrition concerns) and nutrients/food components to encourage (based on undernutrition concerns)
  • For LMICs with micronutrient deficiencies, include relevant micronutrients as encouraging factors or through category-specific adjustments [8]

• Reference Value Alignment:

  • Align threshold values with national or WHO nutrient intake recommendations
  • Adjust reference amounts based on typical consumption patterns (100g, 100kcal, or serving sizes)
  • Validate that thresholds appropriately classify traditionally healthy foods as encouraged

• Expert Consensus Process:

  • Convene multidisciplinary expert panel including nutrition scientists, public health professionals, and food policy experts
  • Use modified Delphi process to assess relevance and appropriateness of model components
  • Document disagreements and resolutions for transparency

Criterion Validation Protocol

Objective: To assess the relationship between foods rated as healthier by the adapted NPM and objective health outcomes.

Materials:

• Prospective cohort or cross-sectional study data with health outcome measures
• Food frequency questionnaires or dietary records
• Statistical analysis software (R, SAS, or STATA)

Procedure:

• Study Design:
  • For prospective designs: Assess association between consumption of foods rated healthier by the NPM and incidence of diet-related diseases
  • For cross-sectional designs: Examine relationships between NPM ratings and biomarkers of nutritional status or disease risk

• Data Analysis:

  • Categorize participants based on diet quality as defined by the NPM
  • Calculate hazard ratios for health outcomes comparing highest and lowest diet quality categories
  • Adjust for potential confounders including age, sex, socioeconomic status, and total energy intake

• Interpretation:

  • Consider the NPM to have substantial criterion validation evidence if consistent, significant associations are observed with multiple relevant health outcomes [6]
  • Compare performance with original NPM framework where possible

Table 2: Key Validation Metrics for Adapted NPMs

Validation Type	Key Metrics	Interpretation Guidelines	Exemplary Performance
Content Validity	Expert consensus on nutrient inclusion	High: ≥80% expert agreement on relevance	PAHO model includes regionally relevant nutrients [11]
Construct/Convergent Validity	Agreement (κ statistic) with reference standard	Near perfect: κ=0.81-1.00; Substantial: κ=0.61-0.80	FSANZ showed near perfect agreement with Ofcom (κ=0.89) [11]
Criterion Validity	Hazard ratios for health outcomes	Significant protective associations for highest vs. lowest diet quality	Nutri-Score associated with lower CVD risk (HR: 0.74) [6]
Discriminatory Ability	Percentage of food supply classified as healthy	Context-dependent; should align with public health goals	NS and HSR classified 22% and 33% of Slovenian food supply as healthy, respectively [62]

Meal-Based Assessment Protocol

Objective: To adapt and validate NPMs for evaluating traditional meals and composite dishes, particularly relevant in cultural contexts with high consumption of homemade meals.

Materials:

• Standardized recipe databases
• Food composition tables
• Meal consumption pattern data

Procedure:

• Model Adaptation:
  • Define typical meal patterns and serving sizes based on local consumption data
  • Select evaluation components addressing public health concerns specific to the population (e.g., protein and vegetables for encouragement; saturated fatty acids and sodium for limitation) [63]
  • Develop scoring algorithm that accounts for meal composition and synergy between ingredients

• Application Testing:

  • Apply the meal-based NPM to diverse meal types commonly consumed in the target population
  • Compare scores across different meal patterns (e.g., traditional vs. modern meals)
  • Assess correlation with existing food-based NPMs and dietary quality indices

• Validation:

  • Test convergent validity against metrics such as Healthy Eating Index-2015 or Nutrient-Rich Food Index [63]
  • Examine association with nutrient density metrics and health outcomes where data are available

The Scientist's Toolkit: Research Reagent Solutions

Table 3: Essential Research Tools for NPM Adaptation and Validation

Tool/Resource	Function	Application Example	Access Considerations
Global Food Monitoring Group Food Categorization	Standardized food categorization system	Enables consistent classification of foods across studies and comparisons between countries [62]	Publicly available with modifications for regional specifics
Composition and Labelling Information System (CLAS)	Branded food composition database	Provides nutritional data for pre-packed foods available in specific markets [62]	Requires collaboration with research institutions or regulatory bodies
Standard Tables of Food Composition	Nutrient composition data for raw and cooked foods	Essential for evaluating traditional foods, recipes, and meals [63]	Country-specific versions available (e.g., Japanese Standard Tables)
Foodbrowser	Commercial food product database	Contains product-specific information including nutritional composition and sales data [63]	Subscription-based commercial resource
Global Health Observatory Data Repository	Population-level nutrition and health indicators	Provides contextual data on malnutrition patterns for situational analysis [8] [59]	Publicly accessible WHO database
Excel Eiyo-kun	Nutritional calculation software	Enables computation of energy and nutrient content of dishes based on ingredient weights [63]	Japanese-specific resource; similar tools exist for other regions

Implementation Considerations and Monitoring Framework

Policy Integration Strategy

Successful implementation of adapted NPMs requires careful consideration of policy contexts and implementation pathways. Front-of-pack labeling represents the most common application of NPMs in LMICs, with specific scheme choices reflecting predominant nutritional concerns [8]. Policy integration should consider:

• Regulatory authority and mandate: Identify appropriate agencies for implementation and enforcement
• Stakeholder engagement: Engage food industry, civil society, and academic partners throughout development and implementation
• Phased implementation: Consider voluntary periods before mandatory implementation to allow for industry adaptation
• Monitoring systems: Establish mechanisms for tracking implementation and impact

Impact Assessment Protocol

Objective: To monitor and evaluate the implementation and impact of adapted NPMs.

Methodology:

• Food Supply Monitoring:
  • Assess reformulation patterns by comparing product compositions before and after implementation
  • Monitor changes in the availability and promotion of healthier products
  • Track the nutritional quality of new product introductions

• Consumer Behavior Assessment:

  • Evaluate changes in consumer awareness, understanding, and use of NPM-informed labels
  • Assess changes in purchasing patterns through household purchase data or sales data
  • Examine potential substitutions and unintended consequences

• Equity Impact Evaluation:

  • Disaggregate impacts by socioeconomic status, geographic location, and other relevant demographic factors
  • Assess differential effects on vulnerable populations
  • Identify potential barriers to equitable benefits

Figure 2: NPM Implementation and Impact Assessment Cycle

Adapting nutrient profiling models for global health applications requires careful balancing of overnutrition and undernutrition agendas, particularly in LMICs experiencing the double burden of malnutrition. The protocols outlined provide a systematic approach for developing contextually appropriate NPMs that respect traditional food systems while addressing modern nutritional challenges. Through rigorous situational analysis, appropriate model selection, comprehensive validation, and careful implementation monitoring, researchers and policymakers can develop effective tools to guide consumers toward healthier choices and stimulate food product reformulation. As global nutrition targets remain elusive, evidence-based adaptation of NPMs represents a promising approach to addressing the complex malnutrition landscape through targeted, context-specific solutions.

The global burden of diet-related non-communicable diseases (NCDs) has established an urgent need for innovative strategies to improve the nutritional quality of the food supply [1]. Nutrient profiling (NP), defined as "the science of classifying or ranking foods according to their nutritional composition for reasons related to preventing disease and promoting health," has emerged as the foundational methodology for guiding these improvements [1]. Within this field, progressive reformulation models represent a transformative approach that enables gradual, incremental product enhancement rather than mandating immediate compliance with the strictest nutritional standards. These models provide a structured pathway for food manufacturers to systematically improve product formulations through a series of achievable steps, balancing technical feasibility with public health objectives [42].

The development of these frameworks responds to the World Health Organization's (WHO) call for food industry engagement in reducing saturated and trans fats, added sugars, and salt in the global food supply [42]. Progressive models address a critical limitation of binary NP systems—their "all-or-nothing" nature—which can inadvertently discourage companies from initiating reformulation efforts when immediate compliance with the highest standards seems technologically challenging or economically prohibitive [42]. By implementing a stepwise improvement strategy, these models create a practical mechanism for continuous nutritional enhancement while maintaining consumer acceptance through gradual taste adaptation.

Established Progressive Nutrient Profiling Models

The PepsiCo Nutrition Criteria (PNC) Framework

The PepsiCo Nutrition Criteria (PNC) represents an internally-used NP model specifically designed to guide and monitor incremental improvements in nutrient density across diverse product categories [42]. This system classifies products into four distinct classes of increasing nutritional value:

• Class IV: Entry-level nutritional quality, establishing a baseline for improvement
• Class III: Meaningful nutritional enhancements implemented
• Class II: Advanced nutritional profile approaching optimal targets
• Class I: Highest nutritional standard, aligned with global dietary recommendations [42]

The PNC model incorporates multiple nutritional dimensions, including nutrients to limit (NTLs), nutrients to encourage (NTEs), and food groups to encourage (FGEs). This comprehensive approach ensures balanced reformulation that not only reduces harmful components but also enhances beneficial nutrients. The system employs category-specific guidelines across twenty distinct product categories, acknowledging that optimal nutrition varies by food type and dietary role [42].

Table 1: Core Nutritional Standards in the PNC Framework

Component Type	Specific Elements	Standard Reference
Nutrients to Limit	Added sugars, saturated fat, sodium, industrially-produced trans fats	≤10% of energy from added sugars and saturated fat; ≤2,000mg sodium daily [42]
Food Groups to Encourage	Fruits, vegetables, whole grains, low-fat dairy, nuts, seeds, pulses, legumes	Minimum ½ serving per portion, with combinations allowed [42]
Gap Nutrients	Country-specific nutrients of public health concern	Based on local dietary guidelines and deficiency patterns

Implementation data demonstrates the model's practical efficacy. An audited review revealed that 48% of products met added sugar goals, 65% met sodium targets, and 71% achieved saturated fat objectives [42]. By the end of 2020, 48% of global beverage sales volume contained 100 kcal or less from added sugars per 355 ml serving, representing 80% of beverage volume and over 90% of food volume sold globally [42].

The Nestlé Nutritional Profiling System

Nestlé employs a comprehensive Nutritional Profiling System that informs product reformulation and development based on four fundamental principles:

• Product category classification
• Identification of specific nutritional factors
• Establishment of thresholds for each nutritional factor
• Reference to individual serving sizes as consumed by adults and/or children [64]

This system emphasizes localized reformulation targets to address specific regional nutritional gaps and accommodate local dietary habits. For example, Nestlé's approach to improving child nutrition differed between the Philippines (adding a daily serving of fortified milk) and Brazil (substituting a daily serving of milk with a fortified version) [64]. This flexibility ensures that reformulation efforts are contextually appropriate and maximally effective.

The integration of advanced technologies enables Nestlé's progressive reformulation, particularly through conversion of sugars to prebiotic fibers and utilization of high-quality proteins with superior Protein Digestibility-Corrected Amino Acid Score (PDCAAS) [64]. This technological dimension distinguishes the model and facilitates meaningful nutritional enhancements that might otherwise be organoleptically challenging.

Food Compass 2.0: An Evidence-Based Profiling System

Food Compass 2.0 represents a recently enhanced nutrient profiling system that evaluates foods and beverages across nine holistic domains of product characteristics, including nutrient ratios, food ingredients, and processing characteristics [5]. This system employs a 100-point scale and incorporates updated scientific evidence on food processing, added sugar, dietary fiber, dairy fat, artificial additives, and trace lipids.

Key advancements in Food Compass 2.0 include:

• Positive points for non-ultraprocessed foods rather than only penalizing ultraprocessed foods
• Better accounting for relatively neutral health effects of dairy fat
• Enhanced recognition of added sugar harms as both additive and food ingredient
• Incorporation of artificial additive data previously unavailable [5]

Validation studies demonstrate that Food Compass 2.0 scores strongly correlate with health outcomes. Each 10.8-point higher individual Food Compass Score (i.FCS) associates with lower body mass index (−0.56 kg/m²), improved blood pressure, favorable cholesterol levels, and reduced prevalence of metabolic syndrome, cardiovascular disease, and cancer [5]. The system also independently predicts all-cause mortality, with a 24% lower risk in the highest i.FCS quintile versus the lowest [5].

Experimental Protocols for Progressive Reformulation

Protocol 1: Product Categorization and Baseline Assessment

Purpose: To establish a systematic baseline of nutritional quality across a product portfolio, enabling targeted reformulation priorities and progress measurement.

Materials and Equipment:

• Laboratory information management system (LIMS)
• Nutritional analysis software (e.g., Food Compass Calculator)
• Chromatographic equipment for nutrient verification [1]
• Standardized nutrient databases

Procedure:

• Categorize products into predefined food categories (e.g., the 20 categories used in PNC) based on dietary role and consumption patterns [42]
• Conduct compositional analysis for key nutrients to limit (saturated fat, added sugars, sodium, trans fats) and nutrients to encourage (protein, fiber, vitamins, minerals) [1]
• Calculate baseline scores using selected NP model (PNC Class, Food Compass Score, or other validated system)
• Establish current distribution across nutritional classes to identify reformulation priorities
• Set incremental targets for migration between classes with defined timelines

Validation Measures:

• Analytical verification of nutrient content claims [1]
• Inter-laboratory comparison for analytical quality control
• Database reconciliation with laboratory results

Protocol 2: Stepwise Reformulation Implementation

Purpose: To systematically improve product formulations through incremental changes that maintain consumer acceptability while enhancing nutritional profile.

Materials and Equipment:

• Ingredient substitution options (e.g., alternative sweeteners, salt replacers)
• Sensory evaluation facilities
• Shelf-life testing equipment
• Pilot-scale production equipment

Procedure:

• Identify reformulation leverage points for target products (e.g., sugar reduction, sodium lowering, saturated fat replacement, fiber enhancement)
• Develop prototype formulations with incremental improvements (e.g., 5-10% reduction in target nutrient per iteration)
• Conduct accelerated stability testing to assess shelf-life implications of reformulation
• Implement sensory evaluation using hedonic scales and difference testing to ensure acceptability
• Validate nutritional changes through analytical testing post-reformulation
• Monitor consumption patterns to assess potential compensatory behaviors [42]

Technical Considerations:

• Ingredient functionality maintenance (texture, mouthfeel, preservation)
• Flavor profile preservation despite reduction of palatability enhancers (sugar, salt, fat)
• Cost implications and supply chain stability of alternative ingredients

Protocol 3: Impact Validation and Monitoring

Purpose: To quantify the public health impact of reformulation efforts and validate associations with health outcomes.

Materials and Equipment:

• Dietary assessment tools (food frequency questionnaires, 24-hour recalls)
• Health outcome biomarkers (cholesterol, blood pressure, glycemic markers)
• Statistical analysis software
• Population consumption data

Procedure:

• Calculate nutritional changes at population level using sales-weighted average nutrient composition [42]
• Model estimated impact on nutrient intakes using dietary modeling approaches
• Validate against health outcomes in observational cohorts when available [5]
• Monitor product portfolio evolution over time using established metrics (e.g., % meeting specific criteria)
• Conduct longitudinal assessment of reformulation impact on consumer preferences

Analytical Methods:

• Chromatographic techniques for micronutrient analysis [1]
• Gas chromatography for fatty acid profiles [1]
• Statistical modeling for health impact projection

Visualization of Progressive Reformulation Framework

The following workflow diagram illustrates the stepwise process for implementing progressive reformulation models:

Progressive Reformulation Workflow

Research Reagent Solutions for Nutrient Profiling

Table 2: Essential Research Reagents and Tools for Nutrient Profiling Studies

Reagent/Tool	Function	Application Example
Gas Chromatography Systems	Separation and analysis of fatty acids, sterols, aroma compounds	Verification of saturated and trans fat reduction in reformulated products [1]
Nutrient Databases	Reference values for nutritional composition	Calculating Food Compass Scores and PNC Class assignments [42] [5]
Sensory Evaluation Tools	Hedonic scales, difference testing protocols	Assessing consumer acceptability of reformulated products [42]
Laboratory Information Management System (LIMS)	Tracking analytical results and reformulation iterations	Maintaining version control and nutritional data integrity [1]
Statistical Analysis Software	Modeling health impacts and consumption patterns	Validating associations between reformulation and health outcomes [5]

Progressive reformulation models represent a pragmatic and effective methodology for systematically enhancing the nutritional quality of food products. The PNC framework, Nestlé Nutritional Profiling System, and Food Compass 2.0 each offer validated approaches for classifying products, establishing incremental improvement targets, and monitoring progress toward public health objectives. The experimental protocols outlined provide researchers with standardized methodologies for implementing and validating these approaches across diverse product categories.

The integration of advanced analytical techniques with sensory evaluation and impact validation creates a comprehensive system for evidence-based reformulation. As food systems worldwide confront the dual challenges of malnutrition and diet-related chronic diseases, these progressive models offer a strategic pathway for aligning food industry practices with public health goals through gradual, achievable improvements rather than disruptive transformations. Future refinement of these models will likely incorporate emerging evidence on nutrient interactions, processing effects, and personalized nutrition responses to further enhance their precision and public health utility.

Leveraging AI and Multi-Omics Data for Next-Generation Dynamic Profiling

The transition from static, population-based nutrient profiling to dynamic, individualized models represents a paradigm shift in nutritional science. This evolution is critical for research comparing traditional and modern crop varieties, as it moves beyond simple compositional analysis to understanding how these foods interact with complex biological systems. Modern multi-omics technologies now enable the comprehensive characterization of molecular responses to dietary interventions, capturing interactions across genomics, transcriptomics, proteomics, metabolomics, and microbiomes [65] [17]. When integrated with artificial intelligence, these data layers facilitate the development of predictive models that can account for individual metabolic heterogeneity, temporal variability in nutritional needs, and personalized physiological responses to specific food components [66] [17].

The convergence of high-throughput omics technologies, microsampling techniques, and advanced computational methods has created unprecedented opportunities for dynamic nutrient profiling. These approaches are particularly valuable for assessing the health impacts of traditional versus modern food varieties, enabling researchers to move beyond static nutrient comparisons to understanding how different food matrices influence metabolic pathways, inflammatory responses, and long-term health outcomes at an individual level [5] [67]. This protocol outlines comprehensive methodologies for implementing AI-driven multi-omics approaches to advance nutrient profiling research.

Foundational Concepts and Current Evidence

Multi-Omics Integration in Biological Research

Multi-omics integration involves the systematic combination of data from multiple biological layers to obtain a comprehensive understanding of health and disease states. The fundamental premise is that each omics layer provides complementary information about biological processes, from genetic blueprint (genomics) to functional phenotype (metabolomics) [65] [68]. In nutritional research, this approach enables the characterization of system-wide responses to dietary interventions, capturing the complex interactions between food components and human biology.

The table below summarizes the core omics technologies relevant to dynamic nutrient profiling:

Table 1: Core Omics Technologies for Dynamic Nutrient Profiling

Omics Layer	Analytical Focus	Key Technologies	Relevance to Nutrient Profiling
Genomics	DNA sequence variations	Whole genome sequencing, SNP arrays	Identifies genetic predispositions affecting nutrient metabolism [65]
Transcriptomics	Gene expression patterns	RNA sequencing, microarrays	Reveals real-time gene regulation in response to food components [65] [68]
Proteomics	Protein expression and modifications	Mass spectrometry, affinity arrays	Quantifies functional effectors of metabolic processes [65]
Metabolomics	Small molecule metabolites	LC-MS, NMR spectroscopy	Captures metabolic endpoints of dietary interventions [65] [67]
Microbiomics	Gut microbiota composition	16S rRNA sequencing, metagenomics	Characterizes food-microbiome interactions [67]

AI and Machine Learning Approaches

Artificial intelligence, particularly machine learning (ML) and deep learning (DL), provides the computational foundation for integrating complex multi-omics datasets. These approaches excel at identifying non-linear patterns across high-dimensional spaces, making them uniquely suited for multi-omics integration [65]. Several architectural approaches have demonstrated particular utility:

• Graph Neural Networks (GNNs): Model biological networks (e.g., protein-protein interactions, metabolic pathways) to identify key regulatory nodes affected by dietary interventions [65] [69]
• Multi-modal Transformers: Fuse disparate data types (e.g., transcriptomic with metabolomic data) to predict integrated responses to nutritional challenges [65]
• Explainable AI (XAI): Techniques like SHapley Additive exPlanations (SHAP) interpret "black box" models, clarifying how specific food components contribute to metabolic outcomes [65]

Recent evidence demonstrates the effectiveness of these integrated approaches. A comprehensive systematic review of dynamic nutrient profiling revealed that AI-enhanced systems demonstrated superior effectiveness (standardized mean difference = 1.67) compared to traditional algorithmic approaches (SMD = 1.08) across multiple nutritional outcome domains [17].

Experimental Protocols

High-Frequency Multi-Omics Microsampling

Principle: Capture dynamic molecular responses to dietary interventions through frequent, low-volume blood collection enabling comprehensive multi-omics profiling from minimal sample volumes [67].

Protocol Steps:

• Sample Collection

  • Collect 10μl blood samples using volumetric absorptive microsampling (VAMS) devices (e.g., Mitra devices)
  • For dietary challenge studies: collect baseline fasted sample, then post-prandial samples at 30, 60, 120, and 240 minutes after test meal consumption
  • Store samples at 4°C during collection day, then transfer to -80°C for long-term storage [67]

• Multi-Omics Extraction

  • Perform biphasic extraction with methyl tert-butyl ether (MTBE) methanol:water (3:1:1, v/v/v)
  • Separate organic phase (lipids, hydrophobic metabolites), aqueous phase (hydrophilic metabolites), and protein pellet
  • Process protein pellet for proteomics analysis using tryptic digestion and LC-MS/MS
  • Perform separate aqueous extraction for multiplexed cytokine/hormone immunoassays [67]

• Data Acquisition

  • Metabolomics: Analyze aqueous and organic phases using UHPLC-MS with reverse-phase and HILIC chromatography
  • Lipidomics: Analyze organic phase using UHPLC-MS with C18 reverse-phase chromatography
  • Proteomics: Analyze digested peptides using nanoLC-MS/MS with data-independent acquisition (DIA)
  • Cytokines/Hormones: Quantify using multiplexed immunoassays (e.g., Luminex platform) [67]

• Data Processing and QC

  • Process raw data using platform-specific pipelines (e.g., Progenesis QI for metabolomics, Spectronaut for proteomics)
  • Apply quality control filters: remove features with >20% missingness in QC samples
  • Normalize data using quality control-based methods (e.g., LOESS, quantile normalization)
  • Annotate metabolites using authentic standards and spectral libraries (e.g., HMDB, LipidMaps) [67]

AI-Driven Multi-Omics Data Integration

Principle: Integrate diverse omics data streams using flexible deep learning frameworks that can accommodate heterogeneous data types and multiple analytical tasks [70].

Protocol Steps:

• Data Preprocessing

  • Perform batch effect correction using ComBat or similar algorithms
  • Impute missing values using appropriate methods (e.g., k-nearest neighbors for metabolomics, missForest for proteomics)
  • Normalize each omics dataset to standard normal distribution (z-scores)
  • Perform feature selection using variance-based filtering and correlation analysis [65] [70]

• Model Architecture Configuration (Using Flexynesis Framework)

  • Select encoder networks based on data characteristics:
    • Fully connected encoders for standard omics data
    • Graph convolutional encoders when incorporating biological network information
  • Configure supervisor multi-layer perceptrons (MLPs) for specific prediction tasks:
    • Regression models for continuous outcomes (e.g., glucose response)
    • Classification models for categorical outcomes (e.g., responder vs. non-responder)
    • Survival models for time-to-event outcomes [70]

• Multi-Task Learning Setup

  • Define multiple supervision heads for simultaneous prediction of related outcomes
  • Implement partial supervision for handling missing labels across different omics modalities
  • Configure loss function weighting to balance contributions from different prediction tasks [70]

• Model Training and Validation

  • Implement stratified train/validation/test splits (typically 70/15/15%)
  • Perform hyperparameter optimization using Bayesian optimization or grid search
  • Apply early stopping with patience of 20-50 epochs to prevent overfitting
  • Validate models using nested cross-validation for robust performance estimation [70]

• Interpretation and Biomarker Discovery

  • Apply explainable AI techniques (SHAP, LIME) to identify feature importance
  • Perform permutation tests to establish statistical significance of identified biomarkers
  • Validate discovered biomarkers in independent cohorts when available [65] [70]

Workflow Visualization

Diagram 1: High-Level Workflow for AI-Driven Dynamic Nutrient Profiling

Diagram 2: AI Architecture for Multi-Omics Data Integration

The Scientist's Toolkit: Essential Research Reagents and Solutions

Table 2: Essential Research Reagents and Computational Tools for Dynamic Nutrient Profiling

Category	Specific Product/Platform	Application/Function	Key Features
Sample Collection	Mitra VAMS Devices (Neoteryx)	Volumetric blood microsampling	Collects precise 10μl samples; suitable for remote collection [67]
Extraction Reagents	MTBE:Methanol:Water (3:1:1)	Biphasic extraction for multi-omics	Simultaneously extracts proteins, metabolites, lipids from single sample [67]
LC-MS Solvents	LC-MS Grade Acetonitrile/Methanol	Mobile phase for chromatography	High purity for sensitive omics detection [67]
Proteomics Standards	iRT Kit (Biognosys)	LC-MS retention time calibration	Enables precise peptide quantification across runs [67]
Metabolomics Standards	MSK-ERC Kit (MS-Omics)	Metabolomics quality control	Standardized quality assessment for longitudinal studies [67]
Computational Framework	Flexynesis (Python)	Deep learning for multi-omics	Modular framework supporting multiple architectures and tasks [70]
Bioinformatics	Galaxy Server Platform	Accessible data analysis	Web-based interface for complex multi-omics workflows [70]
Biological Networks	STRING, KEGG, Reactome	Pathway analysis and network modeling	Curated biological networks for contextualizing omics findings [69]

Analytical Framework for Traditional vs. Modern Varieties

Comparative Multi-Omics Response Profiling

Principle: Characterize differential molecular responses to traditional and modern food varieties through controlled feeding studies with integrated multi-omics profiling.

Protocol Steps:

• Study Design

  • Implement randomized crossover design with washout periods (minimum 1-week)
  • Match traditional and modern varieties for equivalent macronutrient content
  • Include comprehensive phenotyping at baseline (anthropometrics, clinical biomarkers, gut microbiome)
  • Collect multi-omics samples at fasted baseline and multiple post-prandial timepoints [5] [67]

• Data Integration and Analysis

  • Calculate differential response metrics for each omics layer:
    • Metabolomic Response Score: Integrated area under curve for nutritionally relevant metabolites
    • Proteomic Inflammation Index: Dynamic changes in inflammatory cytokines and acute phase proteins
    • Metabolic Flexibility Metric: Rate of return to baseline for energy metabolism markers
  • Apply multivariate statistical methods (PLS-DA, OPLS) to identify discriminatory molecular features
  • Integrate responses across omics layers using multiple kernel learning [5] [17]

• AI-Mediated Personalized Response Classification

  • Train ensemble classifiers to predict responder status for each food variety
  • Identify baseline features (clinical, genomic, microbiomal) predictive of differential responses
  • Develop personalized variety recommendation algorithms based on individual characteristics [65] [17]

Validation and Translation

Principle: Validate discovered nutrient profiles against clinical outcomes and translate findings into practical dietary recommendations.

Protocol Steps:

• Clinical Validation

  • Correlate dynamic molecular profiles with clinically relevant endpoints (glucose homeostasis, lipid metabolism, inflammatory status)
  • Validate prediction models in independent cohorts with diverse demographic characteristics
  • Assess long-term adherence and health outcomes in randomized controlled trials [5] [17]

• Recommendation Engine Development

  • Implement adaptive algorithms that update recommendations based on longitudinal response patterns
  • Incorporate patient preferences, cultural dietary patterns, and accessibility considerations
  • Develop digital interfaces for real-time feedback and dietary adjustment [17]

The integration of these advanced multi-omics and AI methodologies enables a transformative approach to nutrient profiling, moving from static food composition tables to dynamic, biologically relevant models that account for individual variability and system-wide responses. This framework provides the necessary foundation for advancing research on traditional versus modern crop varieties, facilitating evidence-based decisions about agricultural practices, food processing, and personalized nutritional recommendations.

Validating Model Efficacy: Correlating Nutrient Scores with Health and Market Outcomes

Validation is a critical step in ensuring that Nutrient Profiling Models (NPMs) are effective public health tools. Moving beyond theoretical nutritional composition, robust validation frameworks correlate NPM scores with direct measures of health status, including dietary biomarkers in blood and clinical health outcomes. This establishes whether a model accurately reflects the physiological impact of diet and can predict disease risk. Within research on traditional versus modern food varieties, such validation is paramount. It provides an objective, evidence-based means to quantify and compare the health impacts of these distinct food sources, grounding agricultural and nutritional choices in clinical science [71]. This document outlines standardized protocols for conducting this essential validation, aimed at researchers and health scientists.

Experimental Protocols for NPM Validation

Protocol 1: Cross-Sectional Analysis of NPM Scores and Biomarkers

This protocol assesses the relationship between habitual diet quality, as measured by an NPM, and biochemical status.

• Objective: To investigate associations between a dietary index based on an NPM and concentrations of specific blood biomarkers in a study population.
• Key Materials:
  • 24-hour Dietary Recalls: Structured interviews or questionnaires to collect detailed dietary data. Best practice involves multiple recalls (e.g., three non-consecutive days) to account for day-to-day variation [33].
  • Food Composition Database: A comprehensive database to convert consumed foods into nutrient intakes.
  • NPM Algorithm: The specific nutrient profiling model to be validated (e.g., Nutri-Score, PAHO, CAM).
  • Blood Collection Kits: Phlebotomy supplies for collecting fasting blood samples.
  • Analytical Equipment: HPLC, mass spectrometers, or clinical chemistry analyzers for quantifying biomarker concentrations.
• Methodology:
  • Dietary Data Collection & Cleaning: Administer 24-hour dietary recalls to participants. Convert all consumed foods and beverages into nutrient data using the food composition database.
  • NPM Dietary Index Calculation:
    • Calculate the nutrient profiling score for each individual food/beverage item consumed using the NPM algorithm.
    • Compute an overall dietary index for each participant. This is typically done by calculating the weighted average of the NPM scores of all foods consumed, with the proportion of total daily energy intake from each food serving as the weight [33].
  • Biomarker Assessment: Collect fasting blood samples from participants. Process samples (e.g., centrifuge to separate serum/plasma) and analyze for pre-specified biomarkers (e.g., β-carotene, vitamin C, vitamin B9, LDL-cholesterol).
  • Statistical Analysis:
    • Divide participants into groups (e.g., quartiles or quintiles) based on their NPM dietary index.
    • Use Analysis of Variance (ANOVA) to compare mean biomarker concentrations across these groups.
    • Perform trend tests to assess if a significant gradient in biomarker levels exists across increasing levels of diet quality.
    • Use Spearman correlation analysis to evaluate the strength of the relationship between the continuous NPM dietary index and biomarker concentrations [33].

Protocol 2: Validating NPMs Against Food Consumption Data

This protocol evaluates how well an NPM discriminates between different patterns of food consumption, a key aspect of its face validity.

• Objective: To determine if an NPM dietary index can differentiate between consumers of "healthier" and "less healthy" foods, such as traditional versus modern varieties.
• Key Materials:
  • Dietary consumption data (as in Protocol 1).
  • Food Group Classification System: A scheme to categorize foods (e.g., whole grains, refined grains, red meat, fish, fruits, vegetables).
• Methodology:
  • Food Group Aggregation: Classify all consumed food items into pre-defined food groups relevant to the research question (e.g., "whole-grain products," "refined cereal products," "vegetable oils," "plain nuts," "cheeses") [33].
  • Consumption Analysis: Calculate the average daily intake (in grams) of each food group for each participant.
  • Association Testing:
    • Correlate the continuous NPM dietary index with the consumption levels of each food group.
    • Compare food group intakes across quartiles of the NPM dietary index using ANOVA with linear contrasts. A well-validated model should show strong, expected correlations (e.g., higher scores with higher intake of vegetables and whole grains; lower scores with higher intake of sugary drinks and processed meats) [33].

Protocol 3: Model Comparison and Suitability Analysis

This protocol is used to select the most appropriate NPM for a specific population or policy goal by comparing multiple models.

• Objective: To apply and compare various NPMs to a national or regional food supply to determine their suitability for guiding local food policies.
• Key Materials:
  • Food Supply Database: A comprehensive database of packaged foods and beverages available in the target market, including full nutritional composition [72].
  • Multiple NPM Algorithms: The models to be compared (e.g., Nutri-Score, Chilean Warning Octagon (CWO), PAHO, a locally developed model like the Chile Adjusted Model (CAM)) [72].
• Methodology:
  • Data Application: Systematically apply each NPM's algorithm to all products in the food supply database.
  • Categorization: Classify each product as "compliant" or "non-compliant" according to each model's thresholds, or calculate its specific score/rating.
  • Comparison Metrics:
    • Calculate the overall percentage of products classified as "non-compliant" or "less healthy" by each model.
    • Compare non-compliance rates by broad food category (e.g., beverages, dairy, snacks).
    • Analyze key differences in model strictness and the drivers of those differences (e.g., inclusion of free sugars vs. added sugars, handling of non-sugar sweeteners, presence of "positive" nutrients) [72].

The following workflow integrates these protocols into a cohesive validation pipeline:

Key Research Findings and Data

Empirical studies applying these protocols demonstrate the practical validation of NPMs and their relevance to food variety research.

Table 1: Key Findings from NPM Validation Studies

Study / Model	Population / Focus	Validation Method	Key Outcome: Correlation with Biomarkers	Key Outcome: Discrimination of Food Intakes
2023 Nutri-Score Update [33]	French adults (ESTEBAN survey)	Cross-sectional comparison of 2015 vs. 2023 NPM dietary indices with blood biomarkers and food intakes.	Significant inverse trends with β-carotene (Q1=0.76 µmol/L vs. Q5=0.59 µmol/L, p-trend<0.001) and vitamin B9 (Q1=6.68 ng/mL vs. Q5=5.16 ng/mL, p-trend<0.001).	Stronger positive correlations with healthy fats (fish, vegetable oils, nuts) and whole grains. Weaker correlation with cheeses and refined grains than 2015 model.
Chile Adjusted Model (CAM) [72]	South African packaged food supply (n=6474)	Comparison of CAM vs. other NPMs (PAHO, Chilean Warning Octagon) for policy suitability.	N/A - Food supply analysis.	Classified 73.2% of products as non-compliant. Was strictest for beverages (80.4% non-compliance) due to free sugars and non-sugar sweetener criteria.
Traditional vs. Modern Varieties [71]	Global analysis (41 publications)	Review of agronomic, ecological, and cultural services of crop varieties.	Suggests landraces may offer superior nutritional profiles and resilience, highlighting the need for NPMs that can capture these qualities in validation studies.	Landraces often provide more stable yields and regulating services (pest/disease resistance) under sub-optimal conditions, a factor beyond basic nutrient composition.

Table 2: The Scientist's Toolkit: Essential Reagents and Materials for NPM Validation

Item	Function / Relevance in Validation
Standardized Dietary Assessment Tool (e.g., 24-hr recall, FFQ)	To accurately capture habitual food consumption, which is the foundation for calculating the NPM dietary index. Multiple passes and photo aids improve accuracy [33].
National/Regional Food Composition Database	To convert reported food consumption into nutrient intake data. Must be comprehensive and updated to ensure accurate NPM score calculation.
Biomarker Assay Kits (e.g., for β-carotene, vitamin B9, lipids)	To quantitatively measure clinical biomarkers that serve as objective indicators of nutritional status and health outcomes, moving beyond self-reported data [33].
NPM Algorithm Software	A standardized, version-controlled script or software (e.g., in R, Python) to apply the NPM to food consumption data, ensuring reproducibility and transparency.
Statistical Analysis Software (e.g., R, SAS, SPSS)	To perform essential statistical tests like ANOVA, trend tests, and Spearman correlations to establish significant associations between NPM scores and health metrics [33] [72].

Application to Traditional vs. Modern Varieties Research

The comparison of traditional and modern food varieties introduces unique considerations for NPM validation. Modern varieties are often selected for high yield and uniformity, while traditional landraces are valued for resilience and adaptation to local conditions, which can include nutritional quality and density [71]. A robust validation framework must account for these differences.

The following diagram illustrates the hypothesized pathways linking traditional varieties to health outcomes and how NPMs serve as a mediating validation tool:

As shown, NPMs validated against clinical biomarkers provide a crucial quantitative link between the consumption of traditional varieties and potential health benefits. This is especially relevant in Low- and Middle-Income Countries (LMICs) experiencing the double burden of malnutrition, where "Choices" NPM schemes can be designed to encourage micronutrient-dense foods (addressing undernutrition) while limiting sugars, fats, and salt (addressing overnutrition) [8]. Furthermore, when modern eating patterns shift away from traditional foods, validated NPMs can help quantify the health impact of this nutrition transition [73].

Nutrient Profiling Models (NPMs) are quantitative algorithms used to evaluate and rank the healthfulness of foods and beverages based on their nutritional composition [5]. As global strategies to combat diet-related non-communicable diseases intensify, NPMs have become fundamental tools for front-of-pack labeling, marketing restrictions, product reformulation, and public health policy [5] [31]. The performance of these models varies significantly across different food categories, presenting critical challenges for researchers, policymakers, and food manufacturers. This systematic review synthesizes current evidence on NPM performance across diverse food categories, with particular emphasis on the evolving landscape of traditional versus modern food varieties, including plant-based alternatives. Furthermore, it provides detailed experimental protocols for conducting comparative NPM analyses to standardize future research methodologies in this rapidly advancing field.

Results and Comparative Data Analysis

Performance of Updated NPM Systems Across Major Food Categories

Recent developments in NPM methodologies have demonstrated substantial improvements in characterizing food healthfulness. The updated Food Compass 2.0 system, which incorporates emerging evidence on food processing, added sugars, dietary fiber, dairy fat, and artificial additives, shows enhanced capacity to discriminate nutritional quality within and between food categories [5].

Table 1: Food Compass 2.0 Scoring Changes Across Food Categories

Food Category	Original Mean Score	Food Compass 2.0 Mean Score	Percentage Point Change	Key Drivers of Score Change
Seafood	72 ± 14	81 ± 14	+9	Improved accounting of beneficial nutrients
Beef	33 ± 6	44 ± 6	+11	Neutral evaluation of dairy fat, processing attributes
Pork	35 ± 8	44 ± 9	+9	Recognition of minimally processed animal foods
Eggs	46 ± 13	54 ± 13	+8	Processing rather than animal-source considerations
Lamb and Game	39 ± 8	49 ± 8	+10	Beneficial components in minimally processed meats
Rice and Pasta	43 ± 26	49 ± 23	+6	Carbohydrate quality considerations
Cold Cereals	51 ± 21	41 ± 20	-10	Added sugars, processing, artificial additives
Plant-Based Dairy	54 ± 21	43 ± 20	-11	Additives, processing characteristics
Cereal Bars	42 ± 16	34 ± 15	-8	Free sugars, processing, fiber content
Fruit/Veg Juices	72 ± 15	66 ± 14	-6	Free sugars versus total sugars

Comparative Analysis of NPM Applications in Plant-Based Versus Traditional Meat Products

The escalating market for plant-based meat alternatives (PBMAs) has highlighted significant discrepancies in NPM performance when evaluating traditional versus modern food products. A comprehensive analysis of 349 PBMAs using five different classification schemes revealed substantial nutritional variability and model-dependent outcomes [74].

Table 2: NPM Performance in Evaluating Plant-Based vs. Traditional Meat Products

Nutrition Classification Scheme	Type	Performance Characteristics	Key Differentiating Factors
Nutri-Score	Nutrient-based	Effectively identified PBMAs with poor nutritional quality; showed strong agreement with Brazilian NPM for burgers	Energy density, saturated fat, sodium, fiber, protein
Brazilian NPM	Nutrient-based	Recommended as primary driver for PBMA choices due to excellent agreement with Nutri-Score	Specific thresholds for sodium, saturated fat, and inclusion of positive components
PAHO NPM	Nutrient-based	Stringent criteria for nutrients to limit	Focus on free sugars, saturated fat, sodium
NOVA	Food-based	Classified majority of PBMAs as ultra-processed	Degree of processing, presence of additives
Plant-Based Nutrient Profile Model	Hybrid	Combined nutrient and processing considerations	Protein quality, fiber, micronutrient content

Comparative nutritional analysis across European markets revealed that PBM products generally exhibited lower energy density (mean difference: -15.3%), reduced saturated fat content (mean difference: -28.7%), and significantly higher fiber levels (mean difference: +420%) than their meat counterparts [12]. However, protein content remained consistently lower in PBM products (mean difference: -18.9%), while salt levels showed category-dependent variation, with some PBM products exceeding traditional meat equivalents by up to 35% [12].

Impact of NPM Updates on Food Categorization

The evolution from the 2004 UK Nutrient Profiling Model to the proposed 2018 version demonstrates how methodological updates significantly impact product classification outcomes, particularly for specific food categories [31].

Table 3: Impact of UK Nutrient Profiling Model Updates on Product Classification

Food Category	Products Passing 2004 NPM (%)	Products Passing 2018 NPM (%)	Percentage Point Change	Primary Reasons for Change
Beverages	42	11	-31	Switch from total sugars to free sugars
Breakfast Cereals	47	36	-11	Free sugars, fiber thresholds
Yoghurts	62	57	-5	Free sugars definition
Frozen Foods	58	52	-6	Revised nutrient thresholds
Cakes	28	31	+3	Adjustments in scoring algorithm

Experimental Protocols for NPM Comparative Analysis

Protocol 1: Comprehensive Food Product Nutritional Profiling

Purpose: To standardize the collection and analysis of nutritional composition data for comparative NPM evaluation across food categories.

Materials and Reagents:

• Food sampling kits (sterile containers, gloves, labeling materials)
• Nutritional analysis software (e.g., FoodData Central, national composition databases)
• Data extraction forms (electronic or paper-based)
• Statistical analysis package (R, SPSS, or Python with pandas, scipy)

Procedure:

• Product Identification and Categorization
  • Select products using stratified random sampling from major retail chains to ensure market representation
  • Categorize products using standardized classification systems (e.g., FAO/WHO GIFT, USDA Food Patterns)
  • Record brand name, descriptive name, and ingredient list for each product

• Nutritional Data Collection

  • Extract nutritional parameters per 100g/ml: energy (kcal), total fat, saturated fat, unsaturated fat, carbohydrates, total sugars, free sugars, protein, fiber, and sodium/salt
  • For processed foods, additionally record presence of additives: sodium nitrite, monosodium glutamate, natural/artificial colors, preservatives
  • Cross-reference with laboratory analysis where feasible to validate label accuracy

• Data Management and Quality Control

  • Implement double-data entry system with validation checks
  • Establish standardized conversion factors for nutrient interconversion (e.g., sodium to salt: ×2.5)
  • Apply standardized portion size conversion using Reference Amounts Customarily Consumed (RACC)

• Statistical Analysis

  • Conduct descriptive statistics (mean, median, standard deviation) by food category
  • Perform comparative analyses using ANOVA with post-hoc tests for multiple group comparisons
  • Calculate effect sizes for significant differences between traditional and alternative product categories

Protocol 2: Multi-NPM Application and Comparison

Purpose: To evaluate and compare the performance of multiple NPM systems across diverse food categories.

Materials and Reagents:

• NPM calculation tools (algorithm implementation in Excel, R, or Python)
• Comparative analysis framework
• Data visualization software (Tableau, ggplot2, matplotlib)

Procedure:

• NPM Selection and Implementation
  • Select diverse NPM types: nutrient-based (Nutri-Score, Brazil NPM, PAHO NPM), food-based (NOVA), and hybrid models
  • Program exact algorithm specifications for each selected NPM
  • Validate algorithm accuracy using standardized test products with known scores

• NPM Application

  • Apply each NPM to the complete product dataset
  • Record scores and classifications for all products
  • Categorize outcomes according to each model's classification thresholds

• Comparative Analysis

  • Calculate agreement statistics between models using Cohen's Kappa for categorical classifications
  • Assess correlation between continuous scores using Pearson or Spearman correlation coefficients
  • Identify classification discrepancies and outliers for further investigation

• Validation Against Health Outcomes

  • Where possible, correlate NPM classifications with health outcome data from cohort studies
  • Evaluate predictive validity for cardiometabolic risk factors using regression analyses
  • Assess model performance across demographic subgroups to identify equity considerations

Protocol 3: Dietary Pattern-Level NPM Validation

Purpose: To validate NPM performance at the dietary pattern level using food pattern modeling approaches.

Materials and Reagents:

• Dietary intake databases (e.g., NHANES, national nutrition surveys)
• Food pattern modeling software (e.g., USDA Food Pattern Modeling Protocol)
• Nutrient adequacy assessment tools (e.g., Nutrient Adequacy Ratio calculations)

Procedure:

• Dietary Intake Data Preparation
  • Compile individual-level dietary intake data from representative surveys
  • Calculate individual NPM scores (e.g., i.FCS for Food Compass) as energy-weighted averages of foods consumed
  • Categorize participants into quantiles based on dietary NPM scores

• Health Outcome Analysis

  • Assess associations between NPM dietary scores and health parameters using multivariable regression
  • Adjust for potential confounders: age, sex, BMI, physical activity, smoking status, socioeconomic factors
  • Validate against hard endpoints where available: mortality, cardiovascular events, cancer incidence

• Food Pattern Modeling

  • Utilize established USDA Food Pattern Modeling methodologies [75]
  • Evaluate implications of modifying food group quantities within dietary patterns
  • Assess nutrient adequacy when substituting traditional with modern food varieties

Figure 1: Systematic Review Workflow for NPM Comparative Analysis

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 4: Essential Research Reagents and Computational Tools for NPM Research

Category	Item/Resource	Specification/Function	Application Examples
Data Resources	Food and Nutrient Database for Dietary Studies (FNDDS)	USDA database with 8-digit food codes and nutrient profiles	Baseline nutritional composition analysis [76]
	Food Pattern Equivalents Database (FPED)	Converts FNDDS foods to USDA food pattern equivalents	Dietary pattern analysis and modeling [76]
	FoodData Central	USDA's integrated data system on food composition	International comparative analyses [76]
Computational Tools	NPM Algorithm Libraries	Pre-programmed functions for multiple NPM systems	Efficient application of multiple models to product sets [74]
	Statistical Analysis Packages	R, Python, SPSS with specialized nutrition packages	Data analysis, visualization, and modeling
	Food Pattern Modeling Software	USDA-certified modeling protocols	Assessing dietary implications of food substitutions [75]
Methodological Frameworks	PRISMA Guidelines	Systematic review and meta-analysis reporting standards	Ensuring comprehensive and transparent reviews [16]
	Cultural Adaptation Models	PEN-3, Knowledge-to-Action Framework	Adapting NPMs to diverse cultural contexts [43]
	USDA Food Pattern Modeling Protocol	Standardized methodology for dietary pattern analysis	Evaluating impacts of dietary changes on nutrient adequacy [75]

Figure 2: Multi-NPM Application and Validation Workflow

Discussion and Future Research Directions

The comparative analysis of NPM performance across food categories reveals several critical considerations for researchers and policymakers. First, the evolution of NPM systems toward incorporating processing characteristics, free sugars, and specific food ingredients represents a significant advancement in nutritional science [5]. However, this evolution also presents challenges in maintaining consistency for longitudinal research and industry reformulation efforts.

The performance disparities observed between traditional and modern food varieties, particularly in the plant-based meat category, highlight the need for NPM systems that can accurately evaluate both conventional and emerging food products. The substantial variation in nutritional composition within PBMA categories [74] [12] suggests that category-specific approaches may be necessary for meaningful nutritional evaluation.

Future research should prioritize the validation of NPM systems against hard health endpoints across diverse populations, the development of standardized protocols for cross-national NPM comparisons, and the integration of environmental sustainability metrics alongside nutritional considerations. Additionally, as digital nutrition platforms become increasingly prevalent, opportunities exist for integrating dynamic nutrient profiling approaches that incorporate real-time data and artificial intelligence to enhance personalization and accuracy [16] [43].

The methodological protocols presented in this review provide a foundation for standardized, comparable research in this rapidly evolving field, enabling more consistent evaluation of NPM performance across food categories and facilitating evidence-based decisions in public health nutrition policy.

Nutrient profiling models (NPMs) provide quantitative algorithms to evaluate and rank the healthfulness of foods and beverages, serving as critical tools for front-of-pack labeling, marketing restrictions, and public health policies [5]. The evolution of these models reflects an ongoing scientific endeavor to enhance their accuracy in predicting health outcomes. The original Food Compass, developed in 2021, introduced a comprehensive framework assessing 54 attributes across nine health-relevant domains to address limitations in existing systems [18]. Food Compass 2.0 represents a significant advancement, incorporating emerging evidence on food processing, specific ingredients, and their relationships with health [5] [77]. This analysis examines the enhanced predictive validity of Food Compass 2.0 compared to legacy systems within the context of nutrient profiling for traditional versus modern food varieties research.

Comparative Analysis of Nutrient Profiling Systems

Key Advancements in Food Compass 2.0

Food Compass 2.0 maintains the core framework of its predecessor while implementing critical refinements based on the latest scientific evidence. Key updates include: (1) providing positive points for non-ultraprocessed foods rather than only penalizing ultraprocessed foods; (2) better accounting for relatively neutral health effects of dairy fat; (3) enhanced scoring for dietary fiber and long-chain omega-3 fatty acids; (4) more robust evaluation of added sugars as both additives and food ingredients; and (5) incorporation of newly available data on artificial additives [5] [77]. These modifications better align the scoring algorithm with current understanding of how specific food components and processing methods influence health outcomes.

The system evaluates foods on a scale from 1 (least healthy) to 100 (most healthy) based on 54 attributes across nine domains: nutrient ratios, vitamins, minerals, food ingredients, additives, processing, specific lipids, fiber and protein, and phytochemicals [18]. This comprehensive approach allows Food Compass 2.0 to capture nuances often missed by simpler legacy systems, particularly regarding food processing and ingredient quality—crucial dimensions when comparing traditional versus modern food varieties.

Predictive Validity Against Health Outcomes

Table 1: Predictive Validity of Food Compass 2.0 for Health Outcomes in US Adults (n=47,099)

Health Outcome	Association per 1 SD (10.8 points) Higher i.FCS	95% Confidence Interval
Body Mass Index (kg/m²)	-0.56	(-0.65, -0.47)
Systolic Blood Pressure (mm Hg)	-0.55	(-0.77, -0.34)
Diastolic Blood Pressure (mm Hg)	-0.46	(-0.63, -0.29)
LDL Cholesterol (mg/dl)	-1.49	(-2.10, -0.87)
HDL Cholesterol (mg/dl)	+1.61	(+1.41, +1.81)
Total Cholesterol:HDL Ratio	-0.12	(-0.13, -0.10)
Hemoglobin A1c (%)	-0.02	(-0.02, -0.01)
Fasting Plasma Glucose (mg/dl)	-0.36	(-0.67, -0.05)
Metabolic Syndrome (OR)	0.86	(0.83, 0.89)
Cardiovascular Disease (OR)	0.92	(0.88, 0.96)
Cancer (OR)	0.93	(0.89, 0.98)
All-Cause Mortality (HR)	0.92	(0.88, 0.95)

Abbreviations: i.FCS: energy-weighted average Food Compass Score; OR: odds ratio; HR: hazard ratio [5]

Food Compass 2.0 demonstrates robust predictive validity when individual food scores are aggregated to assess overall dietary patterns. In validation studies involving 47,099 U.S. adults, each standard deviation (10.8 points) increase in the energy-weighted average Food Compass Score (i.FCS) was associated with significant improvements in cardiometabolic risk factors and lower prevalence of chronic diseases [5]. The i.FCS showed a high correlation with the Healthy Eating Index-2015 (r=0.78), an established measure of dietary quality [5]. Notably, participants in the highest i.FCS quintile demonstrated a 24% lower risk of all-cause mortality compared to those in the lowest quintile (HR: 0.76; 95% CI: 0.68, 0.84) [5].

Comparative Performance Against Legacy Systems

Table 2: Criterion Validation Evidence for Major Nutrient Profiling Systems

Nutrient Profiling System	Validation Evidence Level	Associated Health Outcomes Validated
Nutri-Score	Substantial	CVD, cancer, all-cause mortality, BMI
Food Compass	Intermediate	Cardiometabolic risk factors, mortality
Health Star Rating	Intermediate	Limited health outcome validation
Nutrient Profiling Scoring Criterion	Intermediate	Limited health outcome validation
Overall Nutrition Quality Index	Intermediate	Limited health outcome validation
Nutrient-Rich Food Index	Intermediate	Limited health outcome validation
Food Standards Agency NPS	Intermediate	Limited health outcome validation

Adapted from systematic review of criterion validation studies [6]

A systematic review of criterion validation studies positioned Food Compass as having intermediate validation evidence, with Nutri-Score demonstrating the most substantial evidence base among major systems [6]. However, direct comparisons reveal Food Compass 2.0's distinctive advantages in several domains. While systems like Health Star Rating (HSR) and Nutri-Score focus primarily on a limited set of nutrients, Food Compass 2.0 incorporates additional dimensions including processing characteristics, specific lipids, phytochemicals, and food additives [5] [78].

Food Compass 2.0 demonstrates enhanced differentiation within food categories compared to legacy systems. For instance, among products receiving the highest HSR (5.0), Food Compass 2.0 scores ranged from 100 (chia seeds) to 10 (fat-free margarine), indicating its superior capacity to discriminate healthfulness among foods that appear similar under simpler systems [5]. Similarly, within NOVA processing group 1 (unprocessed or minimally processed foods), Food Compass 2.0 scores varied from 100 (raw blackberries) to 12 (rice noodles), highlighting its sensitivity to nutritional quality differences beyond processing classification alone [5].

Algorithm Evolution and Validation Pathway of Food Compass 2.0

Application Notes & Experimental Protocols

Protocol for Validating Nutrient Profiling Systems Against Health Outcomes

Objective: To assess and compare the predictive validity of Food Compass 2.0 against legacy nutrient profiling systems for cardiometabolic health outcomes and all-cause mortality.

Materials and Equipment:

• Dietary assessment data (24-hour recalls or food frequency questionnaires)
• Food composition databases
• Health outcome measurements (anthropometric, clinical, mortality data)
• Statistical analysis software (R, SAS, or Stata)

Procedure:

• Food Composition Mapping: Link all food items in dietary assessment data to corresponding Food Compass 2.0 scores and legacy system scores (HSR, Nutri-Score, NOVA)
• Dietary Pattern Scoring: Calculate energy-weighted average scores for each participant:
  • i.FCS = Σ(Food Compass Score × Food Energy) / Total Energy Intake
  • Apply similar calculations for legacy systems
• Health Outcome Assessment: Measure or extract cardiometabolic risk factors (BMI, blood pressure, lipids, glucose metabolism) and disease incidence/mortality data
• Statistical Analysis:
  • Conduct multivariable-adjusted regression models
  • Adjust for age, sex, socioeconomic status, physical activity, smoking, and total energy intake
  • Compare effect sizes per standard deviation increase in each scoring system
  • Assess discrimination using C-statistics or similar metrics
• Validation Metrics: Calculate correlation coefficients with established dietary patterns (e.g., HEI-2015), hazard ratios for mortality outcomes, and odds ratios for disease prevalence

Validation Framework: This protocol was implemented in validation studies involving 47,099 U.S. adults, demonstrating Food Compass 2.0's association with multiple health outcomes after multivariable adjustment [5].

Protocol for Comparing Traditional vs. Modern Food Varieties

Objective: To evaluate differences in healthfulness between traditional and modern food varieties using Food Compass 2.0 and legacy systems.

Materials:

• Traditional and modern varieties of target foods
• Food composition data for both varieties
• Processing classification system (NOVA)
• Food Compass 2.0 algorithm

Procedure:

• Food Selection: Identify paired traditional and modern varieties of specific foods (e.g., whole grain vs. refined grain products; traditional vs. ultraprocessed versions)
• Composition Analysis: Obtain complete nutrient profiles for each food variant
• Systematic Scoring: Apply Food Compass 2.0 and legacy NPMs to all variants
• Processing Classification: Categorize foods using NOVA classification system
• Comparative Analysis: Calculate mean score differences between traditional and modern variants within each system
• Discordance Assessment: Identify instances where systems yield conflicting rankings

Application: This approach enables researchers to quantify how modernization and processing alter the healthfulness of traditional food staples, with Food Compass 2.0 providing more nuanced evaluation through its incorporation of processing and ingredient quality metrics.

The Scientist's Toolkit: Research Reagent Solutions

Table 3: Essential Research Tools for Nutrient Profiling Validation Studies

Research Tool	Function	Application in Food Compass 2.0 Research
NHANES Dietary Data	Nationally representative dietary intake data	Validation of i.FCS against health outcomes in U.S. population
Food Composition Databases	Detailed nutrient profiles for scoring	Implementation of Food Compass algorithm across diverse foods
NOVA Classification System	Food processing categorization	Comparison of processing vs. nutrient-based scoring systems
Healthy Eating Index-2015	Established dietary pattern metric	Criterion validation of Food Compass scoring
STFCJ-8 (Japan)	Country-specific food composition data	Adaptation of profiling systems to different food cultures
Global Health Observatory Data	Population health metrics	Assessment of NPM performance across diverse populations

The validation and application of Food Compass 2.0 requires integration of diverse data sources and methodological approaches. The National Health and Nutrition Examination Survey (NHANES) data provides essential dietary and health information for validation studies, as demonstrated in the initial validation of Food Compass 2.0 with 47,099 U.S. adults [5]. Country-specific food composition databases, such as Japan's Standard Tables of Food Composition in Japan (STFCJ-8), enable adaptation of nutrient profiling systems to local food environments and traditional food varieties [79]. The integration of processing classification systems like NOVA with nutrient-based profiling allows researchers to disentangle the effects of formulation and processing on food healthfulness.

Discussion and Research Implications

Food Compass 2.0 represents a significant advancement in nutrient profiling science through its multidimensional assessment framework and enhanced predictive validity for health outcomes. The system's improved performance stems from its incorporation of emerging evidence on food processing, dairy fats, and specific food ingredients—addressing critical gaps in legacy systems [5] [77]. However, the field continues to evolve, with ongoing research needed to refine scoring algorithms and expand validation across diverse populations.

The application of Food Compass 2.0 to traditional versus modern food varieties research offers particular promise. By simultaneously evaluating nutrient profiles, processing characteristics, and ingredient quality, the system can quantify how food modernization impacts healthfulness in ways that simpler systems cannot capture. This capability is especially relevant for understanding nutrition transitions in low- and middle-income countries, where traditional dietary patterns are rapidly shifting toward processed modern foods [8].

Future research directions should include: (1) validation of Food Compass 2.0 in diverse international contexts; (2) development of simplified scoring methods for practical applications; (3) investigation of cultural adaptations to address regional food traditions; and (4) longitudinal studies examining how changes in Food Compass scores track with health outcome trajectories over time. As nutrient profiling continues to inform public health policies and consumer choices, rigorous validation and refinement of systems like Food Compass 2.0 remains essential for translating nutritional science into effective dietary guidance.

Nutrient Profiling Models (NPMs) are scientific frameworks that classify or rank foods based on their nutritional composition to prevent disease and promote health [80]. Initially developed to guide public health policies, NPMs have evolved into powerful market tools that directly shape the food industry's product development strategies (reformulation) and influence purchasing behaviors (consumer choice). Within the context of research on traditional versus modern food varieties, NPMs provide a quantitative basis for comparing their nutritional quality and health impacts. This document outlines application notes and experimental protocols for analyzing the market impact of NPMs, with a specific focus on their role in product reformulation and their effect on consumer decision-making processes.

Application Notes: The Role and Impact of NPMs

NPMs as a Driver for Product Reformulation

Food product reformulation—altering a product's recipe to improve its nutritional profile—is recognized as a cost-effective intervention for addressing non-communicable diseases, as it can improve population-level diets without requiring significant changes in consumer behavior [81] [82].

• Guiding Reformulation Targets: NPMs provide clear, metrics-based targets for food manufacturers. A model designed for this purpose, such as the Nestlé Nutritional Profiling System (NNPS), is typically category-specific rather than across-the-board. This ensures that improvements are made across all food categories, rather than only in categories that are inherently healthy [82]. For instance, a category-specific model would encourage healthier versions of pizzas and sausages, whereas an across-the-board model might simply discourage their consumption entirely.
• Evidence of Reformulation Impact: Studies demonstrate that systematic reformulation guided by NPMs can significantly improve nutrient intakes at a population level. For example, a review of industry-wide reformulation interventions showed effective reductions in the intake of critical nutrients like sugars, sodium, and saturated fatty acids (SFA) [82].
• Impact on Traditional vs. Modern Food Varieties: The application of NPMs reveals nutritional trade-offs. A 2025 analysis of plant-based meat (PBM) alternatives versus traditional meat products across European markets found that PBMs generally had lower saturated fat and higher fiber, but often contained lower protein and variable salt levels [12]. This underscores the potential for NPMs to guide the improvement of modern food alternatives to better match or exceed the nutritional benefits of their traditional counterparts.

Table 1: Impact of Reformulation Guided by NPMs – Select Examples

Food Category	Reformulation Change	Nutrient/Component Affected	Public Health Implication
Plant-Based Meats [12]	Reduction in saturated fat, addition of fiber	Lower saturated fat, higher fiber	Potential for reduced cardiovascular disease risk
Processed Foods [82]	Reduction in sodium, sugars, SFA	Lower intake of nutrients to limit	Reduced risk of hypertension, obesity, diabetes
Dairy & Beverages [5]	Reduction in added sugars, artificial additives	Improved overall product score	Aligns with dietary guidelines for healthier consumption

NPMs as a Tool for Influencing Consumer Choice

NPMs influence consumer choice primarily through front-of-pack labeling (FOPL), which translates complex nutritional information into an easily understandable format [80] [8].

• Informing and Nudging Consumers: FOPL systems like the Nutri-Score, Health Star Rating, or Warning Labels are operational applications of NPMs. They help consumers quickly identify healthier options at the point of purchase, thereby nudging choices toward products with better nutritional profiles [80].
• Adaptation to Local Nutritional Challenges: The design and implementation of NPMs vary based on regional public health needs. In Latin American countries where overnutrition is prevalent, Warning Label schemes are used to strongly discourage the consumption of energy-dense, nutrient-poor products. In contrast, in Southeast Asia and Zambia where over- and undernutrition coexist, "Choices" FOPL schemes are used. These not only limit negative nutrients but also promote the inclusion of category-specific vitamins and minerals to address micronutrient deficiencies [8].
• Validation of Dietary Patterns: The health impact of consumer choices guided by NPMs can be validated. For instance, the Food Compass 2.0 score was used to calculate an individual's overall diet score (i.FCS). A higher i.FCS was significantly associated with better health outcomes, including lower BMI, improved blood pressure, and reduced prevalence of metabolic syndrome and cardiovascular disease [5]. This provides evidence that consumer choices aligned with high-scoring NPMs correlate with better health.

Table 2: NPM-Based Front-of-Pack Labeling Systems and Consumer Impact

FOPL System Type	Mechanism of Influence	Example Regions	Key Consumer Outcome
Warning Labels	Identifies products high in critical nutrients (e.g., sugar, fat, salt)	Chile, Mexico, Brazil	Effectively reduces selection and purchase of labeled products [8]
Nutri-Score	Color-coded scale from A (green) to E (red)	France, Germany, Spain	Helps consumers compare and choose healthier options within a category [12]
Health Star Rating	Star rating system from ½ (least healthy) to 5 (most healthy)	Australia, New Zealand	Simplifies nutritional comparison across different products [5]
"Keyhole" / "Choices"	Positive logo indicating a healthier choice within a category	North Macedonia, Southeast Asia, Zambia	Encourages selection of products with improved nutritional profiles [8]

Experimental Protocols

Protocol for Analyzing NPM-Driven Product Reformulation

Objective: To quantify the effect of an NPM on the nutritional quality of a product portfolio over time.

Background: Companies often use internal NPMs to set targets for product improvement. This protocol outlines a longitudinal analysis to track reformulation progress.

Materials & Reagents:

• Nutritional Database: Historical nutritional data for target products (e.g., internal product database, public datasets like FNDDS).
• NPM Algorithm: The specific nutrient profiling model used for assessment (e.g., NNPS, Food Compass 2.0).
• Statistical Software: R, Python, or SPSS for data analysis.

Procedure:

• Define Scope and Baseline: Select the product categories and portfolio to be analyzed. Collect comprehensive nutritional data for all products in the scope at the starting time point (T0). This includes energy, macronutrients (protein, carbohydrates, fat, fiber), and micronutrients (sodium, vitamins, minerals) as required by the NPM.
• Calculate Baseline NPM Scores: Apply the NPM algorithm to calculate a score or category (e.g., "Pass"/"Fail") for each product at T0.
• Set Reformulation Targets: Based on the NPM outcomes, define specific, time-bound reformulation targets. For example, "Achieve a 'Pass' score for 80% of products in Category X by T2" or "Reduce average sodium content by 15% in Product Line Y."
• Monitor and Re-score: At predetermined intervals (T1, T2, etc.), recollect nutritional data for the product portfolio. Recalculate the NPM scores for all products.
• Data Analysis:
  • Calculate the percentage of products that have improved their NPM score/category.
  • Track changes in the mean content of key nutrients (e.g., sodium, saturated fat, added sugars, fiber).
  • Perform statistical tests (e.g., paired t-test) to determine if observed changes in nutrient levels and overall scores are significant.

Visual Workflow:

Protocol for Evaluating the Impact of NPMs on Consumer Choice

Objective: To assess how an NPM-based label (e.g., Nutri-Score) influences consumer purchasing behavior in a real or simulated retail environment.

Background: FOPL is a key policy application of NPMs. This protocol uses a controlled experimental design to isolate the effect of the label on consumer choice.

Materials & Reagents:

• Experimental Stimuli: Product images or real products with and without the FOPL application.
• Experimental Environment: Supermarket simulation lab, online store mock-up, or in-store intervention.
• Data Collection Tool: Survey software (e.g., Qualtrics) or point-of-sale data tracking system.
• Participant Recruitment Screener: To recruit a representative sample of the target population.

Procedure:

• Study Design: Choose an appropriate design, such as a randomized controlled trial (RCT) or an A/B testing paradigm.
• Stimulus Preparation: Select a range of products from different categories. Create two sets of stimuli:
  • Control Group: Products displayed with standard nutritional information (no FOPL).
  • Intervention Group: The same products displayed with the FOPL (e.g., Nutri-Score A-E).
• Participant Allocation: Randomly assign participants to either the control or intervention group.
• Task Execution: Present participants with a shopping scenario (e.g., "Select products for a weekly grocery shop"). Ask them to choose from the prepared product sets within a defined budget.
• Data Collection: Record the nutritional profile of the chosen shopping basket. Key metrics include:
  • The average NPM score of the total basket.
  • The proportion of products chosen with "healthy" versus "less healthy" FOPL ratings.
  • Total content of key nutrients (sugar, sat. fat, sodium) in the basket.
• Post-Task Questionnaire: Administer a questionnaire to assess participant awareness, understanding, and perceived influence of the FOPL.
• Data Analysis:
  • Use t-tests or ANOVA to compare the nutritional quality of shopping baskets between the control and intervention groups.
  • Use chi-square tests to compare the frequency of choosing healthier-rated products between groups.
  • Correlate questionnaire responses with actual choice data to understand the drivers of behavior change.

Visual Workflow:

The Scientist's Toolkit: Research Reagent Solutions

Table 3: Essential Tools and Databases for NPM Market Impact Research

Tool / Resource	Type	Primary Function in Research	Example Source / Vendor
Food and Nutrient Database for Dietary Studies (FNDDS)	Database	Provides comprehensive nutritional composition data for foods consumed in the US; essential for calculating NPM scores.	USDA [83]
National Health and Nutrition Examination Survey (NHANES)	Dataset	Links self-reported dietary intake data with health outcomes; used for validating NPMs against mortality and morbidity.	CDC [5] [83]
Nutri-Score Algorithm	NPM Algorithm	A specific FOPL nutrient profiling model; used to test the impact of this particular system on consumer behavior.	Santé Publique France [12]
Food Compass 2.0	NPM Algorithm	A holistic profiling system assessing 9 domains of food characteristics; used for scoring diverse foods, meals, and diets.	Nature Food Journal [5]
Statistical Software (R/Python)	Analysis Tool	Used for data cleaning, calculating NPM scores, performing statistical tests (e.g., Cox models for mortality risk), and generating visualizations.	R Foundation, Python Software Foundation
Supermarket Simulation Software	Experimental Platform	Creates a virtual retail environment for conducting controlled studies on consumer choice and FOPL impact.	Custom-built or commercial platforms (e.g, iMotions)

Nutrient profiling models (NPMs) are quantitative algorithms designed to evaluate and rank the healthfulness of foods and beverages [1]. They have become fundamental tools for public health nutrition policies, including front-of-pack labeling (FOPL), regulation of food marketing to children, and reformulation guidelines [52]. However, many existing NPMs were developed based on nutritional science and dietary patterns prevalent in Western, high-income countries [8]. The direct application of these models to diverse global populations, particularly those with distinct traditional food systems and different burdens of disease, raises critical questions of validity and relevance.

Geographical and cultural validation ensures that NPMs accurately reflect local nutritional priorities, dietary patterns, and public health needs. This process is especially crucial within research comparing traditional and modern food varieties, where standardized models may misclassify culturally significant, nutrient-dense traditional foods. This protocol outlines the systematic procedures for validating and adapting NPMs to ensure their scientific rigor and public health utility across diverse populations.

Key Concepts and Rationale

The Double Burden of Malnutrition

Low- and middle-income countries (LMICs) increasingly face a double burden of malnutrition, where undernutrition (e.g., micronutrient deficiencies) coexists with overnutrition (overweight, obesity, and diet-related non-communicable diseases) [8]. A one-dimensional NPM designed solely to limit nutrients of concern (e.g., sodium, saturated fat, sugars) may inadvertently discourage consumption of energy-dense traditional foods that are vital for addressing undernutrition in certain sub-populations.

Limitations of "One-Size-Fits-All" Models

Systematic reviews of government-endorsed NPMs reveal that their design correlates with a country's predominant nutritional challenges [8]. For instance:

• Warning Label schemes in Latin America primarily discourage consumption of energy-dense, nutrient-poor products where overnutrition is widespread.
• "Choices" schemes in Southeast Asia and Zambia, where over- and undernutrition coexist, incorporate positive messages and encourage the consumption of category-specific vitamins and minerals [8].

This demonstrates that context-specific NPMs are a public health necessity. Furthermore, integrating traditional nutritional wisdom into digital platforms and NPMs is challenging due to standardization barriers, contextual loss, and technological limitations [43].

Experimental Protocols for Geographical and Cultural Validation

This section provides a detailed, step-by-step methodology for validating and adapting an NPM for a specific geographical or cultural context. The process is divided into three sequential phases.

Phase 1: Pre-Validation Contextual Analysis

Objective: To understand the local nutritional, dietary, and cultural context before evaluating or adapting an existing NPM.

• Step 1: Define the Local Public Health Nutrition Landscape

  • Methodology: Conduct a desk review of national health surveys, burden of disease data (e.g., from the Global Burden of Disease study), and government health policy documents.
  • Data Extraction: Quantify the prevalence of key indicators: undernutrition, overweight/obesity, and specific micronutrient deficiencies (e.g., iron, vitamin A, iodine) [8].
  • Output: A summary table of the "double burden" profile for the target population.

• Step 2: Document Local Dietary Patterns and Traditional Foods

  • Methodology: Employ a mixed-methods approach.
    • Cross-Sectional Dietary Survey: Use 24-hour dietary recalls or food frequency questionnaires to collect food intake data from a representative sample [13].
    • Cultural Analysis: Conduct focus group discussions and key informant interviews to document culturally significant foods, traditional preparation methods, and eating rituals [43].
  • Data Analysis: Use multivariate statistical procedures like principal component analysis (PCA) or cluster analysis to identify predominant dietary patterns (e.g., 'traditional', 'western', 'health-conscious') [13].
  • Output: A list of core traditional foods and a characterization of major dietary patterns.

• Step 3: Select and Deconstruct a Candidate NPM

  • Methodology: Choose a well-established NPM (e.g., Food Compass 2.0, Nutri-Score, Health Star Rating) as a candidate for adaptation [5] [52]. Deconstruct its algorithm to understand:
    • Nutrients and Components: All included nutrients (to limit and to encourage) and food components (e.g., fiber, whole grains, food processing attributes) [52].
    • Scoring Algorithm: Base unit (per 100g vs. per 100kcal), weighting of components, and threshold values.
    • Underlying Health Rationale: The evidence base used to develop the model.

Phase 2: Model Evaluation and Adaptation

Objective: To test the candidate NPM against local criteria and modify it to improve its relevance.

• Step 4: Model Testing and Quantitative Gap Analysis

  • Methodology: Apply the candidate NPM to a comprehensive local food composition database, including documented traditional foods.
  • Data Analysis:
    • Identify traditional foods that are classified as "unhealthy" by the candidate model but are considered nutrient-dense or culturally important locally.
    • Flag modern, ultra-processed foods that are classified as "healthy" by the model but conflict with local dietary guidelines.
  • Output: A quantitative gap analysis table highlighting misalignments.

• Step 5: Stakeholder Delphi Panel for Model Adaptation

  • Methodology: Convene a panel of local experts (nutrition scientists, public health policymakers, clinical dietitians, cultural anthropologists, and consumer representatives). Use a modified Delphi technique with iterative rounds of anonymous voting and feedback to achieve consensus on necessary model changes [43].
  • Adaptation Proposals: The panel may propose:
    • Adjusting nutrient thresholds (e.g., for sodium or saturated fat) to account for traditional preservation methods.
    • Adding or modifying components to reflect local micronutrient deficiencies (e.g., promoting iron- or vitamin A-rich traditional foods) [8].
    • Incorporating positive points for specific traditional food groups (e.g., legumes, nuts, fermented foods) or for being "non-ultraprocessed" [5].
  • Output: A culturally adapted NPM algorithm.

Phase 3: Validation and Implementation

Objective: To validate the adapted NPM against health outcomes and develop implementation protocols.

• Step 6: Health Outcome Validation

  • Methodology: In a cohort or cross-sectional study, calculate the energy-weighted average NPM score for each participant's diet. Validate this score against objective health outcomes using multivariable regression analysis, adjusting for confounders like age, sex, and physical activity [5].
  • Health Markers: Body mass index (BMI), blood pressure, lipid profiles, blood glucose levels, and prevalence of metabolic syndrome [5].
  • Output: A report on the association between the adapted NPM score and health outcomes, confirming its predictive validity in the local context.

• Step 7: Implementation Protocol Development

  • Methodology: Based on the validated model, develop context-appropriate implementation tools. This could include:
    • Culturally Adapted Front-of-Pack Labels: Using participatory design with the target community to ensure icons and messages are understood [43].
    • Digital Integration: Incorporating the adapted NPM into digital nutrition platforms using flexible data architectures and culturally-informed algorithms to make it accessible to the public [43].

The following workflow diagram illustrates the comprehensive validation protocol:

Data Presentation and Analysis

The following tables summarize the quantitative and qualitative data essential for the validation process.

Table 1: Prevalence of Nutritional Challenges Influencing NPM Design in Selected LMICs (Conceptual Data)

Region/Country	Undernutrition Prevalence	Overweight/Obesity Prevalence	Common Micronutrient Deficiencies	NPM Response Strategy
Latin America	Lower	High	Varies	Warning Label: Focus on limiting sugar, fat, salt [8]
Southeast Asia	Coexists	Coexists	Iron, Vitamin A	"Choices" Scheme: Limit negative nutrients; promote category-specific vitamins/minerals [8]
Zambia	Coexists	Coexists	Varies	"Choices" Scheme: Limit negative nutrients; promote category-specific vitamins/minerals [8]
North Macedonia	Lower	High	Varies	"Keyhole" FOPL: Limit sugar, fat, salt; promote fibers, fruits, vegetables [8]

Table 2: Comparison of Nutrient Profiling Model Characteristics

Model (Example)	Key Nutrients to Limit	Key Nutrients/Components to Promote	Scoring Base	Unique Features / Contextual Considerations
Food Compass 2.0 [5]	Added sugar, sodium, saturated fat, artificial additives	Fiber, whole grains, fruits, vegetables, nuts, legumes, seafood, non-ultraprocessed foods	per 100 kcal	Comprehensive (9 domains); positive points for minimally processed foods; validated against health outcomes.
PAHO Model [8] [1]	Sodium, free sugars, total fats, saturated fats	(Varies by application)	per 100 g	Targets processed and ultra-processed foods; used for various public health strategies in the Americas.
Health Star Rating (HSR) [52]	Sodium, saturated fat, total sugars	Protein, fiber, fruit/vegetable/nut/legume content	per 100 g	Algorithm varies by food category; used for front-of-pack labeling in Australia and New Zealand.
"Choices" Schemes [8]	Sodium, saturated fat, total sugars, trans fat	Category-specific micronutrients (e.g., iron, calcium, vitamins)	(Varies)	Designed for contexts where over- and undernutrition coexist; encourages fortification.

The Scientist's Toolkit: Research Reagent Solutions

Table 3: Essential Materials and Tools for Geographical and Cultural Validation Research

Item / Reagent	Function / Application in Validation Research	Specific Examples / Notes
Food Composition Database	Provides nutrient data for local and traditional foods required for model scoring.	Must be locally compiled or expanded to include traditional foods and dishes (e.g., Argentine locro, Japanese natto) not found in international databases [43].
Dietary Assessment Software	Enables collection and analysis of dietary intake data for pattern identification and model validation.	Software like Globodiet or tools integrated with local food databases are essential for accurate 24-hour recalls [13].
Statistical Analysis Software (e.g., R, Stata)	Performs multivariate statistical analysis (PCA, cluster analysis) and regression modeling for health outcome validation.	Used to derive dietary patterns from food intake data and to test associations between NPM scores and health biomarkers [13] [5].
Cultural Adaptation Frameworks	Provides a theoretical structure for respectfully translating and integrating traditional knowledge.	Frameworks like Bernal's model (adapting language, persons, metaphors, content, concepts, goals, methods, context) guide the ethical adaptation process [43].
WebAIM Color Contrast Checker	Ensures that any visual outputs (e.g., FOPL icons, dashboard elements) meet accessibility standards (WCAG).	Critical for implementation, ensuring graphical elements have a 3:1 contrast ratio and text has a 4.5:1 ratio for inclusivity [84].

Geographical and cultural validation is not an optional enhancement but a fundamental requirement for the ethical and effective application of nutrient profiling models in global research and policy. The rigorous, multi-phase protocol outlined herein—encompassing contextual analysis, participatory model adaptation, and health-outcome validation—provides a roadmap for developing NPMs that are both scientifically sound and culturally respectful. For research focusing on traditional versus modern food varieties, this process is indispensable. It ensures that models accurately classify and valorize nutrient-dense traditional foods, thereby supporting public health goals, preserving cultural heritage, and promoting sustainable dietary patterns tailored to the unique needs of diverse populations worldwide.

Conclusion

The effective application of nutrient profiling models is paramount for advancing research at the intersection of nutrition, drug development, and public health. A one-size-fits-all approach is insufficient; instead, a nuanced strategy is required. This involves selecting or developing NPMs with careful consideration of the public health context—whether addressing overnutrition or micronutrient deficiencies—and ensuring they are culturally and technically adapted for both traditional and modern food varieties. Future progress hinges on standardizing dynamic profiling methodologies, integrating multi-omics data for truly personalized nutrition, and conducting long-term validation studies. For biomedical researchers, embracing these advanced, ethically-grounded profiling tools will be critical for developing targeted nutritional interventions, validating health claims, and formulating functional foods that meet the diverse needs of global populations.

References

• 1. An overview of nutritional profiling in foods - PubMed Central [https://pmc.ncbi.nlm.nih.gov/articles/PMC10070841/]
• 2. Nutrient Profile Models with Applications in Government-Led ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC6247226/]
• 3. Editorial: Adaptation of traditional crop cultivars to climate ... [https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2024.1427068/full]
• 4. Nutrient density: principles and evaluation tools [https://pubmed.ncbi.nlm.nih.gov/24646818/]
• 5. Food Compass 2.0 is an improved nutrient profiling system ... [https://www.nature.com/articles/s43016-024-01053-3]
• 6. Criterion validation of nutrient profiling systems [https://www.sciencedirect.com/science/article/pii/S0002916523661878]
• 7. How to Develop Nutrient Profiling Models Intended for ... [https://www.sciencedirect.com/science/article/pii/S2161831322001247]
• 8. Nutrient Profiling Models in Low- and Middle-Income ... [https://www.sciencedirect.com/science/article/pii/S2475299124024648]
• 9. Nutrient Profiling - an overview [https://www.sciencedirect.com/topics/food-science/nutrient-profiling]
• 10. A Comparative Analysis of Graphic Models for Enhancing ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12195978/]
• 11. Comparison of nutrient profiling models for assessing the ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC6137431/]
• 12. Comparative analysis of the nutritional quality of plant ... [https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2025.1603600/full]
• 13. Traditional v. modern dietary patterns among a population in ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10260468/]
• 14. Associations between the overall nutritional quality of ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12590596/]
• 15. The Nutrient Rich Foods Index helps to identify healthy ... [https://www.sciencedirect.com/science/article/pii/S0002916523017847]
• 16. A Comprehensive Systematic Review of Dynamic Nutrient ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12610451/]
• 17. A Comprehensive Systematic Review of Dynamic Nutrient ... [https://www.mdpi.com/2304-8158/14/21/3625]
• 18. The Tufts Food Compass [https://tuftsfoodismedicine.org/project/food-compass/]
• 19. Comparison of Nutri-Score and Health Star Rating Nutrient ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10002453/]
• 20. Updated Food Compass 2.0 improves food health ratings ... [https://www.news-medical.net/news/20241013/Updated-Food-Compass-20-improves-food-health-ratings-using-new-diet-data.aspx]
• 21. New Rating System Aims to Demystify Nutrition [https://www.foodmanufacturing.com/ingredients/news/22670913/new-rating-system-aims-to-demystify-nutrition]
• 22. Data – The Tufts Food Compass [https://sites.tufts.edu/foodcompass/research/data/]
• 23. Methods – The Tufts Food Compass [https://sites.tufts.edu/foodcompass/research/methods/]
• 24. Food Quality Scores and Nutrient Ratios Correlate with ... [https://www.medrxiv.org/content/10.1101/2024.12.23.24319328v1.full]
• 25. Perspective: How to Develop Nutrient Profiling Models ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC8166553/]
• 26. Why Traditional Food Knowledge Matters in Modern ... [https://www.tdu.edu.in/post/why-traditional-food-knowledge-matters-in-modern-nutrition-science]
• 27. The Vital Role of Traditional Foods in Achieving Nutrition ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10706084/]
• 28. Facilitators and barriers to the consumption of traditional ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC8857890/]
• 29. Statistical Methods for the Analysis of Food Composition ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9182527/]
• 30. Food for health: Putting traditional foods back on the table [https://bhma.org/food-for-health-putting-traditional-foods-back-on-the-table/]
• 31. The Nutrient Profiling Model explained: Stay ahead of ... [https://www.igd.com/articles/the-nutrient-profiling-model-explained-stay-ahead-of-hfss-changes/71452]
• 32. Restricting the promotion of high fat, salt and sugar foods ... [https://www.gov.wales/restricting-promotion-high-fat-salt-and-sugar-foods-implementation-guidance-annex-4-html]
• 33. Impact of the 2023 revision of the nutrient profiling system ... [https://pubmed.ncbi.nlm.nih.gov/40481832/]
• 34. National food tables vs brand-level nutrient data [https://www.myfood24.org/blog/national-food-tables-vs-brand-level-data/]
• 35. USDA FoodData Central [https://fdc.nal.usda.gov/]
• 36. Food and Nutrient Database for Dietary Studies 2021‐2023 [https://www.sciencedirect.com/science/article/abs/pii/S0889157525009822]
• 37. Updated FDA Guidelines for 2025 Including “Healthy” Food ... [https://foodmedcenter.org/updated-fda-guidelines-for-2025-including-healthy-food-labeling-how-they-compare-to-the-2016-guidelines-and-what-it-could-mean-for-health-consumers-and-manufacturers/]
• 38. An overview of nutritional profiling in foods: Bioanalytical ... [https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2023.1124409/full]
• 39. Should nutrient be ‘ profile ' or... models category specific [https://pmc.ncbi.nlm.nih.gov/articles/PMC2879603/]
• 40. (PDF) Nutrient schemes: overview and profiling analysis comparative [https://www.academia.edu/4332077/Nutrient_profiling_schemes_overview_and_comparative_analysis]
• 41. Should nutrient be ' profile ' or... models category specific [https://www.ndph.ox.ac.uk/food-ncd/publications/162821]
• 42. A Progressive Nutrient Profiling System to Guide ... [https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2021.774409/full]
• 43. Integrating Traditional Nutritional Wisdom into Digital Nutrition ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12196328/]
• 44. Comparative analysis of the nutritional quality of plant-based ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12307488/]
• 45. Toward Nutritionally Sound Plant-Based Meat Analogues [https://www.mdpi.com/2304-8158/14/17/3068]
• 46. Healthiness of Meat-Based Products in Comparison to ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11544961/]
• 47. Plant-based meat vs real meat | Differences explained for ... [https://thesustainableagency.com/blog/plant-based-meat-vs-real-meat/]
• 48. Article 8(j) - Traditional Knowledge, Innovations and ... [https://www.cbd.int/traditionalknowledge]
• 49. Genetic Resources, Traditional Knowledge and ... [https://www.wipo.int/en/web/traditional-knowledge]
• 50. Launch of ''Indigenous Knowledge, Ancestral Places [https://www.unesco.org/en/articles/launch-indigenous-knowledge-ancestral-places-navigating-change-unesco-designated-sites]
• 51. Development of the Ajinomoto Group Nutrient Profiling ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12133537/]
• 52. Systematic Review of Nutrient Profile Models Developed ... [https://www.sciencedirect.com/science/article/pii/S2161831323013637]
• 53. How to choose colors for data visualizations [https://www.atlassian.com/data/charts/how-to-choose-colors-data-visualization]
• 54. Defining the role of processing in food classification ... [https://www.nature.com/articles/s41538-025-00395-x]
• 55. The Emerging Role of Ultra-Processed Foods in Obesity ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11976848/]
• 56. Why Is the American Diet So Deadly? [https://www.newyorker.com/magazine/2025/01/13/why-is-the-american-diet-so-deadly]
• 57. Nutrient profiling and degree of food processing of child ... [https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0330687]
• 58. Crediting Traditional Indigenous Foods in Child Nutrition ... [https://www.fns.usda.gov/cn/crediting-traditional-indigenous-foods]
• 59. Nutrient Profiling Models in Low- and Middle-Income ... [https://pubmed.ncbi.nlm.nih.gov/39886080/]
• 60. Assessing progress towards the global nutrition targets [https://globalnutritionreport.org/reports/2021-global-nutrition-report/assessing-progress-towards-the-global-nutrition-targets/]
• 61. The nutrient - GOV.UK profiling model [https://www.gov.uk/government/publications/the-nutrient-profiling-model]
• 62. Comparison of Nutri-Score and Health Star Rating Nutrient ... [https://www.mdpi.com/1660-4601/20/5/3980]
• 63. Frontiers | Development of the Ajinomoto Group Nutrient ... Profiling [https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2025.1568181/full]
• 64. Nestlé drives product reformulation based on nutrient density [https://www.foodnavigator-asia.com/Article/2025/07/17/nestle-drives-product-reformulation-based-on-nutrient-density/]
• 65. AI-driven multi-omics integration in precision oncology [https://pmc.ncbi.nlm.nih.gov/articles/PMC12634751/]
• 66. AI-driven multi-omics integration for multi-scale predictive ... [https://www.sciencedirect.com/science/article/pii/S2001037024004513]
• 67. Multi-omics microsampling for the profiling of lifestyle- ... [https://www.nature.com/articles/s41551-022-00999-8]
• 68. Advances in Integrated Multi-omics Analysis for Drug ... [https://www.mdpi.com/2218-273X/14/6/692]
• 69. Network-based multi-omics integrative analysis methods in ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11954193/]
• 70. Flexynesis: A deep learning toolkit for bulk multi-omics data ... [https://www.nature.com/articles/s41467-025-63688-5]
• 71. More than Yield: Ecosystem Services of Traditional versus ... [https://www.mdpi.com/2071-1050/10/8/2834]
• 72. Applying and comparing various nutrient profiling models ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9378746/]
• 73. Understanding traditional and modern eating: the TEP10 ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC6889524/]
• 74. Nutrition classification schemes for plant-based meat ... [https://www.sciencedirect.com/science/article/pii/S2665927124001229]
• 75. Food Pattern Modeling for the 2025 Advisory Committee ... [https://www.dietaryguidelines.gov/2025-advisory-committee-report/food-pattern-modeling]
• 76. Development of Ferrero Nutrition Criteria: category-specific ... [https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2025.1612183/full]
• 77. Food Compass 2.0 Updates Food Healthfulness Rankings [https://solutions.tufts.edu/expertise/food-nutrition/food-compass-20-updates-food-healthfulness-rankings]
• 78. Tufts University's Food Compass: Is It 'Broken'? [https://optimisingnutrition.com/tufts-food-compass/]
• 79. Development of a Nutrient Profiling Model for Processed ... [https://www.mdpi.com/2072-6643/16/17/3026]
• 80. Nutrient profile models a useful tool to facilitate healthier ... [https://www.sciencedirect.com/science/article/abs/pii/S0924224421000911]
• 81. Nutrient profiling for product reformulation: public health ... [https://pubmed.ncbi.nlm.nih.gov/28420455/]
• 82. Nutrient profiling for product reformulation: public health ... [https://www.cambridge.org/core/journals/proceedings-of-the-nutrition-society/article/nutrient-profiling-for-product-reformulation-public-health-impact-and-benefits-for-the-consumer/CDCD9DC269CA4292DA20E5B795F37A01]
• 83. Development and Validation of a Nutrient Profiling Model ... [https://www.medrxiv.org/content/10.1101/2025.01.03.25319947v3.full-text]
• 84. Data visualisation: colours - Government Analysis Function [https://analysisfunction.civilservice.gov.uk/policy-store/data-visualisation-colours-in-charts/]