Training Perspectives from the NIH T32 Artificial Intelligence for Precision Nutrition Programs

Abstract

Nutrition is one of the few modifiable risk factors for health status. The diet-disease relationship, however, is highly complex, with many endogenous and exogenous factors contributing to risk and with variation in responses to diet between individuals. The 2020-30 National Institutes of Health (NIH) Strategic Plan for Nutrition Research states that precision nutrition is a unifying and holistic approach to developing comprehensive and dynamic nutritional recommendations relevant to both individual and population health. Recognizing the growing importance of advanced computational approaches in nutrition research, in 2022 the NIH Office of Nutrition Research and participating NIH Institutes and Centers published a Funding Opportunity Announcement for training programs in AI for Precision Nutrition (AIPrN) aiming to develop a new generation of the scientific workforce, literate in both nutritional and computer sciences, that leverages the growing data resources available for nutrition research to tackle complex and emerging health-disease associations. In this paper, we describe the approaches developed by the first four institutional awardees of the T32 AIPrN training grant, and we highlight applied strategies for interdisciplinary training, data integration, and methodological innovation as the research community prepares to analyze and interpret data generated by major initiatives such as the NIH-funded Nutrition for Precision Health Initiative.

Background

Food and nutrition influences every individual, every day, and has profound effects on cognitive, physical, and social well-being. Nutritional deficits directly account for nearly 50 million disability-adjusted life years (DALYs) globally (1). In addition, they are also one of the major risk factors for most of the leading causes of morbidity and mortality in the United States (US), and abroad per Global Burden of Disease (GBD) data (2). The interplay between nutrition and health is well established; poor nutrition affects both disease risk and severity whereas the disease can in turn further worsening nutritional status, creating a bidirectional relationship (3–6). While this was articulated by Nevin Scrimshaw almost 50 years ago (7), we still today recognize nutrition as one of the few truly ‘modifiable risk factors’ for health status. In fact, the 2020 GBD analysis suggest nutrition-associated factors (e.g. healthy eating, diet diversity, maintaining a healthy weight, normal blood pressure and fasting glucose) and physical inactivity as the largest contributors to death and disability across both high and low-resource settings (2). Specifically, approximately one-third of all premature deaths in the US, including those from non-communicable diseases, such as diabetes and obesity, are attributable to suboptimal diets (8). The global numbers are equally concerning with deaths attributable to poor diets having grown by 15% since 2010, outpacing population growth, accounting for 26% of all adult deaths in a year (9).

Despite the critical effect of diet on health, the diet-disease relationship is highly complex. Disease risk is influenced by many endogenous (e.g. genetics, epigenetics, and microbiome composition) and exogenous factors (e.g. diet, physical activity, smoking) and individuals often exhibit highly variable responses to the same dietary exposures (10–12). Consequently, uniform dietary recommendations may not adequately address the diverse nutritional needs of different populations. This challenge is underscored by the continuously high prevalence of nutrition-related chronic conditions. In fact, nearly two-thirds of the US population are diagnosed with obesity or overweight (13), and it is also estimated that globally one in every three individuals has one or more micronutrient inadequacies (both deficiencies and excesses, such as in sodium) (14). Notably, this number can be as high as 50% depending on the setting (15, 16). Recognizing these challenges, the US government and the academic community have recently taken steps to shift focus to precision nutrition efforts through initiatives such as the Nutrition for Precision Health (NPH) program. NPH is the first ancillary study of the All of Us Research Program and is the hallmark of the broader shift of nutrition research toward large-scale, population biobank data to better understand individual variability in dietary responses and inform personalized dietary recommendations (17).

Central to these efforts is understanding the concept of precision nutrition itself. Precision nutrition leverages individual-level multilayer data, such as multiomic (metabolomic, genomic, proteomic, metagenomic), diet, behavioral and environmental, to provide individualized dietary guidance designed to optimize health outcomes (18). The 2020-30 National Institutes of Health (NIH) Strategic Plan for Nutrition Research expand on this concept by envisioning precision nutrition research as the unifying and holistic framework for developing comprehensive, targeted, effective and dynamic nutritional recommendations for both individuals and populations (19). A recent publication reporting on the proceedings from the NIH’s January 2021 online workshop on “Precision Nutrition: Research Gaps and Opportunities” underscores the complexity of precision nutrition research and highlights the amount of data that is needed to support it (20).

In response to the continuously growing complexity and scale of available data, precision nutrition research is increasingly adopting advanced Artificial Intelligence (AI) and Machine Learning (ML) approaches to integrate, analyze, and interpret dietary and health data (21). These approaches can identify complex patterns within large, heterogeneous datasets and capture relationships that are often nonlinear, high-dimensional, and influenced by continuously changing environmental interactions, presenting unique challenges in both data analysis and interpretation (22, 23). At the same time, major data sources are continuously expanding in both the US and globally, including the NIH-supported All of Us Researcher Workbench (17), the Common Fund Data Ecosystem (24), and Biodata Catalyst (25), as well as global initiatives such as the GBD datasets and large consortia focused on microbiome and multi-omics (26). Together, these platforms offer unique and unprecedented opportunities to study complex associations of nutrition, biology, and health-disease mechanisms.

Recognizing the growing importance of AI-enabled precision nutrition research, in 2022 the NIH Office of Nutrition Research (ONR) and participating NIH Institutes and Centers (ICs) published a Funding Opportunity Announcement (FOA) (27, 28) for institutional training programs in AI for Precision Nutrition (AIPrN) at the predoctoral and/or postdoctoral level, emphasizing integration across precision nutrition, AI/ML approaches, systems biology and systems science, “big data,” and computational analytics. Aiming at bringing together faculty from nutrition, medicine, and related health and biomedical sciences, along with trainers from computer, engineering and data science, the proposed program will develop a new generation of the scientific workforce that leverages the growing data resources available for nutrition research to tackle complex and emerging health-disease associations.

In this work, we describe the approaches developed by the first four institutional awardees of the T32 AIPrN training grant, and we highlight applied strategies for interdisciplinary training, data integration, and methodological innovation as the research community prepares to analyze and interpret data generated by major initiatives such as the NIH-funded NPH. The remainder of the paper is organized as follows. We first provide an overview of each training program, including program structure and administrative organization across the four institutions awarded the T32 AIPrN training grant (Cornell University, Ithaca, New York; Prairie View Agri & Mech (A&M) University, Prairie View, Texas; the University of Hawai‘i at Mānoa, Honolulu, Hawaii and the College of Health Solutions at Arizona State University, Phoenix, Arizona) followed by a description of the implementation of the AIPrN training program at each site. We then present our approaches to training, with a focus on curriculum development, mentorship models, and the integration of artificial intelligence and precision nutrition methodologies.

Cornell University: Artificial Intelligence and Precision Nutrition Training Program

Saurabh Mehta (Contact PI); David Erickson (MPI); Deborah Estrin (MPI), Julia L. Finkelstein (MPI); Thorsten Joachims (MPI), Rulla M. Tamimi (MPI); Diana M. Thomas (MPI); Martin T. Wells (MPI).

Program Overview

The Cornell AIPrN training program represents the first institutional training grant focused on the intersection of AI and nutrition and is supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), NIH. The training program is designed to train the next generation of scientists and build a workforce equipped with expertise in AI and ML methods to tackle complex biomedical challenges in nutrition and health using high-dimensional data. The program brings together faculty from Cornell University’s Ithaca, Cornell Tech, and Weill Cornell Medicine campuses, West Point, and University of California San Diego from the fields of nutrition, medicine, health and biomedical sciences, computer and information sciences, data science, engineering, and bioinformatics with core competencies in computational methodology, data science, systems science, ML, AI, and data infrastructure. The program is directed by Professor Saurabh Mehta; co-directors include Professors Julia Finkelstein (Cornell University), Deborah Estrin (Cornell Tech), Thorsten Joachims (Bowers CIS), Martin Wells (Bowers CIS), Rulla Tamimi (Weill Cornell Medicine), David Erickson (Engineering), and Diana Thomas (USMA West Point).

To bridge disciplines and strengthen interdisciplinary and translational impact, the program proposes a comprehensive approach to AI-driven nutrition research. Trainees are guided from the earliest stages of defining biologically meaningful and nutrition relevant questions, through data evaluation and preparation, model development, validation, reproducible workflow design and ultimately real-world application (Table 1). This approach ensures that trainees are prepared to develop skills in modern AI methods that extend beyond basic programming and apply them in ways that are scientifically robust, and meaningful for both nutrition and information sciences. This is well aligned with NIH priorities in precision nutrition which emphasize rigorous dietary assessment, strong analytics, and real-world impact across the lifespan (19).

Table 1.

Integrated Framework Connecting Project Stages with Core Competencies, Datasets, and Deliverables for AI-Enabled Precision Nutrition Training

Phase	Key questions	AIPrN training components	Core competencies Nutrition and Health Domains	Core competencies CS/Data Domains	Datasets	Deliverables
Project understanding	What exactly is the problem and expected benefit? What would a solution look like? What is known about the domain?	Nutrition foundations (NS 6310 Micronutrients, NS 6320 Macronutrient Metabolism)	Frame biologically-based question Identify modifiable risk factors Identify diet-disease pathways Design public-health and clinical studies	Translate health domain-related questions into AI/ML problems Define success criteria and constraints Clarify requirements	All of Us, NPH, cohort/clinical data, HMRU human studies, maternal-child health/cancer projects	Problem and Specific Aims statement NS 7030 presentations proposal pitch to Special Committee
		Epidemiology and translational research (NS 6520, NS 6200) Seminar (NS 7030)
		Special Committee mentoring and monthly PD meetings;
		Focus areas: Maternal-Child Health and Cancer
Data understanding	What data are available? Are they relevant, valid, and representative? Do they reflect expectations? Are quality, quantity, and recency sufficient?	eCornell certificate in AI and Precision Nutrition BTRY 6010/6020 Data seminars rotations/externships across DNS, Weill Cornell PHS, and CIS	Interpret dietary assessments, biomarkers, multi-omics Evaluate measurement error and bias Assess clinical relevance	Data auditing-profiling EDA for tabular, text, image, time-series Interoperability (e.g., FHIR/HER Data governance and privacy	EHR/INSIGHT CRN, microbiome/metabolomics, wearables/sensors, nutrition surveys (ASA24,)	Data audit and data dictionary embedded in capstone workflow Feasibility/scoping memo
Data preparation	Which data should we concentrate on? How is the data best transformed for modeling? How may we improve data quality?	Reproducible pipelines in eCornell certificate in AI and Precision Nutrition	Harmonize dietary, clinical, and-omics data control for confounding Handle longitudinal structure	Feature engineering batch correction, imputation, class-imbalance handling Workflow automation (R/Python)	Linked multi-modal cohorts cross-institutional HER NPH modules Lab assay QC	Versioned pipeline (code + README) Data provenance and QC report Reproducibility checklist
		Statistics/Data Science coursework DNS/CIS joint mentoring
		Ontology mapping of foods, biomarkers, and clinical codes
Modeling	What model family/architecture fits the problem? What is the best technique/method? How good is the model technically?	CS 4700/5700 Foundations of AI	Apply ML to maternal-child and cancer problems Interpret outputs with biological plausibility Derive clinically meaningful endpoints	Model selection and tuning DL (CNNs/transformers) Causal ML uncertainty and calibration Fairness-aware ML	Risk prediction, phenotype clustering, dietary pattern mining, digital twins	Model training notebook Performance dashboard Model card and datasheet
		Dual-mentor research projects
		CS 4780/5780 Intro to ML
		CS 5726
		CS 4740/5740 NLP
		CS 4769/5670 CV
		Advanced ML (CS 6780/6785/6789)
Evaluation	How good is the model with respect to project requirements? What have we learned?	eCornell certificate in AI and Precision Nutrition course projects	Benchmark against clinical standards/guidelines Assess generalizability across populations and settings	Robust validation (CV, external validation), calibration Explainability (SHAP/LIME) Error analysis	Multi-site EHR contexts via WCM collaborations Subgroup/equity analyses when data permit	Capstone evaluation section Advisory feedback annual progress Report + IDP Candidacy/disser tation milestones
		BIOMG 7510 ethics and professional responsibilities
		Program PI/Advisory Board review
		Internal-external validation activities
Deployment	How is the model best deployed? How do we know the model is still valid?	Leadership and translation (ILRHR 7451)	Translate findings into guidelines/interventions Consider feasibility, access, and equity	Packaging and deployment (APIs/apps) Monitoring and model drift Documentation and governance	Clinical decision support and public-health translation settings accessed through WCM and Cornell units	Dissemination products (talks, manuscripts) Proposal submissions (postdocs: K01/F32/K99/R01)
		Grant writing (NS 7040)
		NS 7030 dissemination
		Tech translation via Engineering/Cornell Tech
		Weill Cornell implementation science and decision support exposure

Each project phase is linked to the training components, core competencies, datasets, and deliverables required for AI-enabled Precision Nutrition (AIPrN) program. It follows the structure of the data-mining lifecycle known as the Cross-Industry Standard Process for Data Mining-(CRISP-DM)(33) tailored to nutrition science. Abbreviations: AIPrN, Artificial Intelligence for Precision Nutrition; AoU, All of Us Research Program; CV, computer vision; DL/ML, deep learning / machine learning; DNS/CIS, Division of Nutritional Sciences / Computer & Information Science; EHR, Electronic Health Record; IDP, Individual Development Plan; INSIGHT CRN, INSIGHT Clinical Research Network; NLP, natural language processing; NPH, Nutrition for Precision Health; QC, quality control; WCM, Weill Cornell Medicine;.

Cornell has 14 colleges and schools across Ithaca, New York City, and Doha campuses. With nearly 100 fields of study, Cornell’s unique system of cross-disciplinary and integrative graduate study is, however, independent of traditional departments, colleges or campuses and has the kind of multidisciplinary training required by the FOA as a feature. Cornell leverages its existing strengths and initiatives in implementing this novel program to train the next generation of scientists in the domains of AI and precision nutrition to address the future needs for the nutrition workforce. The training program, with positions for 4 predoctoral and 2 postdoctoral trainees per year, is built on the outstanding doctoral programs in the multiple departments at Cornell University participating in this application, which emphasizes multidisciplinary and integrative scholarship across the biological, natural, behavioral, data, and social sciences. The program’s multidisciplinary structure is contributing to reducing the isolation often experienced by trainees working across different fields, particularly those without direct access to data science expertise (29). Specifically, the 23 trainers that are part of the T32 in AIPrN represent the broad range of disciplines necessary to achieve the goals of the training program and include renowned scientists with expertise spanning from nutrition, genetics, molecular biology and biochemistry to epidemiology, bioinformatics, psychology, and economics. The program is not only providing state-of-the-art data science-centered training but also structured mentorship and guidance to help students navigate through the continuously rapidly developing domain of AI (29). The trainers have active research programs and strong records of mentorship which include leading large-scale cohort studies and randomized trials in maternal and child nutrition, integrating genetic, molecular, and lifestyle data to study obesity and cancer risk, developing diagnostic and sensing technologies, and advancing ML methods for biomedical and population health applications. Collectively, the trainers have mentored well over one hundred pre- and postdoctoral trainees, many of whom have pursued research-focused careers in academia, industry, and the government. Further, the cross-cutting natures of the AI Initiative that a T32 program MPI leads and is the newly appointed Vice Provost for (TJ) facilitates the implementation (30).

Program Administration

Cornell’s AIPrN training program is led by a multi-PI team comprising seven faculty members representing the departments/units at Cornell with the necessary expertise, trainer pool, and eligible trainees to achieve the goals of this funding opportunity (Figure 1).

Figure 1.

Cornell University AIPrN training program organization

The Training and External Partnerships Core at the newly launched Cornell Joan Klein Jacobs Center for Precision Nutrition and Health in the College of Human Ecology serves as the institutional home for the T32 activities and the scholars. The Jacobs Center and the Division of Nutritional Science’s Office of Administrative Support provide administrative assistance, financial accounting, and reporting for the program. All trainers are contributing to the proposed program by (a) advising the multi-PI team on program operation, trainee selection, and evaluation of trainee progress; (b) serve as ambassadors of the training program and help with advertising and recruiting to the program in their individual departments; and (c) as members of the Special Committee, or as mentors who supervise the individual programs of both predoctoral and postdoctoral trainees. The Special Committee is the group that guides and evaluates the progress of PhD candidates at Cornell and is selected by the trainee in consultation with the chair of the committee to best achieve their given training goals, described in greater detail below.

All predoctoral trainees in this training program have two co-chairs from the nutrition and health track and from the computing and information science track each. All postdoctoral trainees are co-mentored similarly. The dissertation research for predocs and research project(s) for postdocs is focused on AI and precision nutrition.

In addition to formal coursework offered across participating units, the Cornell program has developed a new, fully online eCertificate in Artificial Intelligence for Precision Nutrition. The eCertificate has been designed to teach trainees the foundations of data science, reproducibility, ethical AI use, and core computational methods tailored for nutrition research with practical applications in the nutrition domain. In addition, a series of hands-on workshops and training seminars designed specifically for the T32 trainees from all four institutions, such as the intensive methods workshop held in September 2025, an in-person workshop in May 2026 hosted in New York City, and a satellite trainee-focused workshop planned during the 2026 American Society for Nutrition (ASN) meeting. Together, these activities provide a structured learning platform for cross-institutional training and create opportunities for trainees and mentors from different sites and AIPrN programs to collaborate, learn together and present ongoing work.

Prairie View Agri & Mech University: Increasing Future AI-Precision Nutrition Research Workforce to Promote Nutrition and Health

Seungchan Kim (Contact PI), Janet Antwi (MPI), Salma Musaad (MPI)

Program Overview

The overarching goal of the AIPrN training program at Prairie View A&M University (PVAMU) is to train a cadre of predoctoral students from Electrical and Computer Engineering and postdoctoral Nutrition Science scholars to become future leaders in AIPrN. The AIPrN training program at PVAMU in partnership with the USDA/ARS Children’s Nutritional Research Center (CNRC) at the Baylor College of Medicine is designed to provide multidisciplinary training in AI and ML and precision nutrition to address the unbalanced access to adequate nutrition and health based on real-world research projects covering food insecurity, obesity, type 2 diabetes, cardiovascular diseases, and cancer.

PVAMU, a Historical Black College/University (HBCU) located in Prairie View, Texas, serves as the lead and host institution. As part of the Texas A&M University System, PVAMU is a land grant university authorized under the Morrill Acts of 1890 and is accredited by the Southern Association of Colleges and Schools Commission on Colleges, Inc with a large enrollment (97%) of students with limited resources, needing additional training opportunities. PVAMU is located in Waller County, which is designated as a rural and food desert area with low-income and health-care desert neighborhoods (31).

Each year, the program aims to train three predoctoral students housed in Electrical and Computer Engineering and one postdoctoral scholar housed in Nutrition Science at PVAMU. The program combines rigorous didactic courses in precision nutrition and AI/ML, research methods, ethics, mentorship, and professional and career development. The program is enhanced by strong AI/ML research in Electrical and Computer Engineering, a well-established Nutrition Science program addressing precision nutrition and health differences at PVAMU in partnership with the Institute for Advancing Health through Agriculture (IHA) at Texas A&M AgriLife Research, and cutting-edge research at the USDA/ARS CNRC at Baylor College of Medicine in Texas Medical Center. The program assembles an interdisciplinary team of 16 faculty mentors from 5 Departments, 4 Colleges, and 2 Institutions (PVAMU-Electrical and Computer Engineering, Agriculture, Nutrition & Human Ecology, Biology, Health & Kinesiology; BCM-CNRC, Graduate School of Biomedical Sciences) with broad range of research expertise in precision nutrition, computational systems biology, AI/ML, genomics, obesity, type 2 diabetes, cancer, microbiology, advancing healthcare access, and eating disorders.

Program Administration

The cross-disciplinary AIPrN training program is co-led by a multi-PI/PD team, namely, Dr. Seungchan Kim, Dr. Janet Antwi, and Dr. Salma Musaad. The 3 PDs/PIs work closely with the faculty mentors as the Executive Committee (Figure 2. Administrative Structure below). The postdoctoral trainee that is recruited into the program every year acts as the representative of the trainees on the Executive Committee to maintain a direct line of communication between trainees and the leadership team. There is an Internal Advisory Committee that consists of the Deans of the College of Agriculture, Food and Natural Resources, and the College of Engineering at PVAMU, and the External Advisory Committee includes the Director of CNRC, Director of Postdoctoral Fellowship Program of CNRC, and the Director of the IHA Texas A&M AgriLife Research. The Advisory Committees advise the Executive Committee on effective monitoring and assessment of the program both internally and externally for effective project execution. The T32 training program is also closely working with the PVAMU Graduate School for program monitoring and support. There is also an Administrative Support person to coordinate events and support oversight activities between the departments, colleges, and institutions involved, trainee-related activities, and liaison among the trainees, Program Faculty, and Executive Committee. All predoctoral trainees complete core courses, including: ELEG 636 Modern Artificial Intelligence, ELEG 6316 Statistical Learning for Big Data, GNEG 5315 Data Ethics, NUTR 5304 Fundamentals of Integrated AI-Precision Nutrition, NUTR 5310 Nutrition Assessment, ELEG 4377 Machine Learning for Engineering Applications, ELEG 6316 Statistical Learning for Big Data, GNEG 5315 Data Ethics, NUTR 5324 Research Seminars in AI-Precision Nutrition, NUTR 5328 Advanced Research Methods, Statistics and Data Analysis in Nutrition, and NUTR 5329 Analytical methods for health equity research in nutrition and health (32). Elective courses include for example, ELEG 6318 Deep Learning, NUTR 5312 Social Cultural Influences on Nutrition, and GS-IY-6205 Microbiome Methodology & Data Analysis.

Figure 2.

Administrative structure of the AIPrN Research Training Program at Prairie View A&M University (PVAMU). Abbreviations: CAHS, College of Agriculture and Human Sciences; COE, College of Engineering, CNRC; Children’s Nutrition Research Center; PVAMU, Prairie View A&M University; TAMU AgLife; Texas A&M AgriLife. The dual mentoring arrow references that each trainee has two primary faculty mentors.

University of Hawai‘i at Manoa: The Hawaii Advanced Training in Artificial Intelligence and Nutrition Science Research

Youping Deng (Contact PI), Rachel Novotny (MPI)

The University of Hawai‘i at Manoa (UHM) is a minority serving land grant institution with established success fostering a diverse group of next-generation biomedical and nutrition investigators. This mixed intercollege predoctoral and postdoctoral training program leverages the cross-disciplinary excellence in research conducted by the applicants with diverse backgrounds and expertise in various connected fields, such as nutrition science, and nutrition-related chronic diseases, children’s health, as well as systems biology and big data analytics, large-scale bioinformatics and omics (genomics, epigenomics, proteomics, metabolomics, microbiomes) accompanied with robust computational approaches, e.g., AI and ML. The selected trainees aim to make significant discoveries from large datasets in nutrition and metabolic diseases, cancer, diabetes, cardiovascular diseases, and health disparities. Distinct from the many training programs, AIPrN trainees have two primary mentors who have expertise in one of the two areas: one in nutrition or nutrition-relevant biomedical sciences, and the other in AI including ML, computational or data science analytic approaches such as engineering, computer science, applied mathematics, or statistics.

An important strength is collaboration among UHM faculty in the John A. Burns School of Medicine (JABSOM), the College of Tropical Agriculture and Human Resilience (CTAHR), UHCC, the Pacific Biosciences Research Center (PBRC), the Information & Computer Sciences Department, and the neighbor healthcare medical centers, in a shared commitment to improving health and well-being for Hawaii’s diverse population. The AIPrN Training Program is timely for leveraging synergies with the UHM Centers of Biomedical Research Excellence (COBRE) research centers including the Pacific Center for Artificial Intelligence and Data Science in Medicine (PAC-AID), the Diabetes, Cardiovascular, and Biogenesis (developmental and reproductive biology), Infectious Diseases, and Microbiomes, and Precision Nutrition in Human Health, the Children’s Healthy Living Center of Excellence (CHL Center), as well as the local Kuakini Medical Center COBRE Center for Translation Research on Aging (CTRA), all addressing nutrition and nutrition-related health and diseases. The NIMHD Research Centers at Minority Institutions (RCMI) Ola Heath And Wellness Achieved by Impacting Inequalities (HAWAII) and Pacific Innovations, Knowledge, and Opportunities (PIKO) training programs in the UHM offer additional synergies and the potential to leverage various institutional resources, including Investigator Development, Research Infrastructure, Community Engagement and Regulatory Knowledge and Support Cores.

However, none of the current training programs at UHM describe cross-disciplinary training specifying advanced AIPrN as areas of research training focus. The Hawaii program has 33 faculty with broad expertise in the two key domains of precision nutrition research with AI/ML and advanced analytics. We have had an outstanding diverse trainee pool in the past 10 years that are eligible for support. Our program faculty has research programs that fit well to the program objectives to train next generation leaders in precision nutrition research. This Hawaii AIPrN program is the vibrant research and training hub for AI-Powered Precision Nutrition and nutrition-related disease and health research. Core courses in the Hawaii program are: Fundamentals of Data Science, Bioinformatics, Machine Learning for Data Science, Nutrition and Disease, and Advanced Food Science. Elective courses include: Biostatistics, Data Management and visualization for health sciences, Secondary data analysis, Topics in Nutritional Sciences, Nutritional Epidemiology, DNA and Genetic analysis, Information theory in Machine Learning, Deep Learning with neural networks, advanced Artificial Intelligence and Advance applied health disparities research. Career development focused courses include: Responsible conduct of research, scientific writing, and grant writing.

Arizona State University Tempe Campus: Interdisciplinary Systems-based Training for Precision Nutrition

Corrie Whisner (Contact PI), Li Liu (MPI)

Program Overview

The INterdisciplinary sysTEms-based tRAining in preCision nuTrition (INTERACT) program at Arizona State University (ASU), supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) of the NIH as part of the AIPrN Initiative, was established to prepare a new generation of researchers capable of integrating nutrition and metabolism science with advanced AI and systems modeling approaches. Rooted in ASU’s transdisciplinary academic structure, the program is intentionally designed to cultivate scientists who can work fluidly across biological, computational, clinical, and community contexts to address the complex determinants of diet-related chronic disease.

INTERACT emphasizes a systems-oriented approach in which AI methods are developed within biological, behavioral, and environmental contexts. Trainees work with multiscale data spanning molecular to population-level factors to generate clinically meaningful and interpretable insights. This training ensures that computational models reflect underlying physiology and real-world complexity, enhancing the rigor and translational relevance of AI-driven discoveries.

Recognizing that precision nutrition research relies on complex data derived from diverse populations, INTERACT integrates training in ethical data stewardship, community engagement, and reciprocity. Trainees learn principles of responsible data governance, including transparency, participant autonomy, and engagement of underserved and historically marginalized communities as research partners. Through targeted learning modules, trainees develop practical skills to design studies that incorporate community input, ensure appropriate data use, and promote bidirectional benefits.

The program brings together faculty across nutrition science, metabolism, biomedical informatics, engineering, and data science within ASU’s transdisciplinary College of Health Solutions (CHS). This structure fosters cross-disciplinary collaboration and provides a rich environment for integrative, translational training.

INTERACT offers a structured yet flexible curriculum combining foundational coursework with individualized training plans. Trainees begin with a week-long Precision Nutrition and Artificial Intelligence Introductory Bootcamp, followed by individualized leveling coursework in nutrition (NTR 510: Nutrition Across the Lifecourse) or analytics, depending on prior training. Foundational coursework includes Research Methods, Machine Learning for Health Applications, and Grant Writing, while trainees complete an interdisciplinary Systems-based Modeling for Precision Nutrition course developed specifically for the program. Elective coursework spans nutritional biochemistry and metabolism (e.g., Advanced Micronutrient Metabolism, Advanced Macronutrient Metabolism, Nutrigenomics, Medical Nutrition in the Care and Prevention of Disease) together with AI and data science (e.g., Statistical Learning for Data Mining, Translational Bioinformatics, Bayesian Statistics, Deep Learning in Visual Computing, and Sensors and Machine Learning). Scientific seminars, journal clubs, and individualized research experiences further support training. Together, these components prepare trainees to analyze complex health data, apply AI and systems modeling tools, and translate findings into actionable hypotheses and interventions. In parallel, trainees will complete modules in data stewardship, ethical AI, and community-engaged research practices.

Each trainee is co-mentored by faculty in nutrition/metabolism and one in AI/systems modeling, ensuring integration of perspectives across disciplines (Figure 3). Training includes cross-disciplinary research practicums, experiential learning in clinical, community, or industry settings, including engagement with community partners, and opportunities for leadership and scientific communication. These experiences prepare trainees to connect computational discoveries with biological meaning and societal impact.

Figure 3.

Overview of the INterdisciplinary sysTEms-based tRAining in preCision nuTrition (INTERACT) training program, Arizona State University.

Through its systems-based approach, INTERACT prepares researchers to advance precision nutrition in diverse, real-world settings and to lead the next generation of AI-enabled nutrition research and practice.

Program Administration

The INTERACT program employs a shared leadership model reflecting the interdisciplinary nature of precision nutrition and artificial intelligence. The program is co-directed by Dr. Corrie Whisner, who oversees the nutrition and metabolism components of the program, and Dr. Li Liu, who oversees the artificial intelligence and systems modeling components. Together, the co-directors coordinate trainee recruitment, curriculum implementation, mentor selection, program evaluation, and strategic planning to ensure integration across the complementary scientific domains.

Program operations are managed through an Executive Committee composed of the co-directors, the directors of the Exercise and Nutritional Sciences and Biomedical Informatics PhD programs, and the Assistant Dean for Graduate Education within the College of Health Solutions. The committee meets regularly to oversee trainee recruitment, mentor matching, curriculum implementation, financial administration, and evaluation of trainee and mentor progress while fostering an inclusive and collaborative training environment. A Steering Committee consisting of faculty leaders from relevant academic and administrative units provides institutional guidance on program development, integration of university-wide training resources, and long-term strategic priorities.

An External Advisory Committee composed of leaders from partner healthcare systems and research organizations, including the Phoenix Veterans Affairs Healthcare System, the NIDDK Phoenix Epidemiology and Clinical Research Branch, Valleywise Health, and Mayo Clinic, provides independent evaluation of program progress and helps ensure that training remains responsive to emerging clinical, community, and research needs. Day-to-day program activities are coordinated by a dedicated program coordinator who supports trainee recruitment, seminar and workshop organization, mentor communication, professional development activities, budget management, and maintenance of program resources. Together, this governance structure provides continuous oversight, facilitates interdisciplinary collaboration, and supports a cohesive trainee experience across the program.

Discussion, Future Steps and Coordination Plan

As of now, the four programs share similar objectives, including the integration of nutrition related health and computational sciences, and the training of predoctoral and postdoctoral trainees from varied backgrounds. Although, the programs have largely operated independently, we aim to strengthen the coordination across them, beginning with this joint Perspective paper. By fostering collaboration, we intend to achieve a national strategic priority (19), while addressing a global need for the multidisciplinary training of future health scientists who will be able to integrate and contribute to both nutrition and computational sciences. To achieve this, we will establish a coordinated cross-site framework that spans the entire data science workflow, from study design and data collection to data harmonization and integration, data analysis and interpretation, and ultimately the translation and dissemination of findings (Table 1). First, we will develop shared training resources aligned with the stages of this workflow. These will include courses available online for asynchronous use, ML model development focusing on translation of findings into policy and practice. All modules will be developed with the contributions of all institutions and delivered in formats that accommodate differences in trainee backgrounds. Second, we will organize joint, hands-on analytic training sessions and workshops using Common Fund datasets (All of Us, NPH). Outputs such as code, workflows, and documentation will be maintained in shared repositories to promote reproducibility and reuse. Third, we will foster ongoing collaboration across the four programs through preconference workshops at flagship meetings, such as the annual meeting of the American Society for Nutrition, and networking events. In addition, another area of focus is recruitment, retention, and career development of the current trainees who will have the opportunity to present their research as an example of the kind of work that can be done leveraging the program’s interdisciplinary framework. In conclusion, the NIH-supported AIPrN Training Program is the first coordinated effort to build a national workforce at the intersection of AI and precision nutrition, recognizing the potential role of technology to transform public health. This cross-university coordination model provides a practical and scalable approach to sustaining precision nutrition training.

Bringing together expertise across institutions and disciplines, the program prepares trainees to apply advanced data science methods to increasingly complex, multimodal datasets relevant to nutrition and health. In addition, by leveraging existing institutional strengths and initiatives, AIPrN is the first training program that will train the next generation of scientists in the domains of both AI and nutrition equipping them with the necessary skills to leverage the growing availability of data resources and use them to tackle complex biomedical challenges in nutrition science. Program graduates are anticipated to go into academia, government, as well as industry, and lead independent programs and/or projects in precision nutrition using the tools gained from this training program. Through learning from peer programs, sharing modules and analytic pipelines, and convening cross-site exchanges, AIPrN presents an interconnected training ecosystem that will lead to the development of innovative solutions to combat diet-related chronic diseases and nutrition disparities within the mission areas of the participating NIH Institutes and Offices. We have shared this perspective on the benefits and challenges of this unique type of cross disciplinary training program to hopefully serve as a roadmap to guide the next generation of similar programs.

Abbreviations:

AI

Artificial Intelligence

AIPrN

Artificial Intelligence for Precision Nutrition

A&M

Agri & Mech

AoU

All of Us

ARS

Agricultural Research Service

ASN

American Society for Nutrition

ASU

Arizona State University

CHL

Children’s Healthy Living Center of Excellence

CHS

College of Health Solutions

CIS

Computing & Information Science

CNRC

Children’s Nutritional Research Center

COBRE

Centers of Biomedical Research Excellence

CTAHR

College of Tropical Agriculture and Human Resilience

CTRA

Center for Translation Research on Aging

CV

computer vision

DALY

Disability-adjusted life year(s)

DL

Deep learning

DNS/CIS

Division of Nutritional Sciences

EHR

Electronic Health Record

FOA

Funding Opportunity Announcement

GBD

Global Burden of Disease

HAWAII

Heath And Wellness Achieved by Impacting Inequalities

ICs

Institutes and Centers

IDP

Individual Development Plan

INSIGHT CRN

INSIGHT Clinical Research Network

INTERACT

INterdisciplinary sysTEms-based tRAining in preCision nuTrition

IHA

Institute for Advancing Health

JABSOM

John A. Burns School of Medicine

ML

Machine Learning

MPI

Multiple Principal Investigator(s)

NICHD

National Institute of Child Health and Human Development

NIDDK

National Institute of Diabetes and Digestive and Kidney Diseases

NIH

National Institutes of Health

NIMHD

National Institute on Minority Health and Health Disparities

NLP

Natural language processing

NPH

Nutrition for Precision Health

PAC-AID

Pacific Center for Artificial Intelligence and Data Science in Medicine

PBRC

Pacific Biosciences Research Center

PD

Project Director

PhD

Doctorate of philosophy

PI

Principal Investigator

PIKO

Pacific Innovations, Knowledge, and Opportunities

PVAMU

Prairie View A&M University

QC

Quality control

RCMI

Research Centers at Minority Institutions

SDI

Socio-demographic Index

UHM

University of Hawaii at Manoa

US

United States

USDA

United States Department of Agriculture

USMA

United States Military Academy

WaSH

Water, Sanitation, and Hygiene

WCM

Weill Cornell Medicine