Exploration of biomarkers of food intake in a Caribbean Hispanic population

Authors: Parnell, Laurence; FOUHY, LIAM - University Of Massachusetts, Lowell; AKINLAWON, OLADIMEJI - ARS Postdoctoral Research Associate; Lai, Chao Qiang; NUSETOR, FREDERICK - Tufts University; ORDOVAS, JOSE - Tufts University; MANGANO, KELSEY - University Of Massachusetts, Lowell; TUCKER, KATHERINE - University Of Massachusetts, Lowell; NOEL, SABRINA - University Of Massachusetts, Lowell

Submitted to: Molecular Nutrition and Food Research
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/13/2025
Publication Date: 7/21/2025
Citation: Parnell, L.D., Fouhy, L.E., Akinlawon, O., Lai, C., Nusetor, F., Ordovas, J.M., Mangano, K.M., Tucker, K.L., Noel, S.E. 2025. Exploration of biomarkers of food intake in a Caribbean Hispanic population. Molecular Nutrition and Food Research. e70158. https://doi.org/10.1002/mnfr.70158.
DOI: https://doi.org/10.1002/mnfr.70158

Interpretive Summary: Certain small molecules commonly found in blood offer a potential solution as unbiased indicators of consumed food. Thus, a set of such validated indicators was explored in a Caribbean Hispanic population in order to gauge their usefulness in assessing the intake of specific foods. Several such molecular indicators were observed. Additionally, the research uncovered a number of common genetic variants that associate with some of these molecular indicators of food intake. Together, the findings imply that the accuracy of some blood molecules in representing the intake of specific foods may depend partly on a person’s genetics. This information is useful for developing advanced mathematical models, which integrate genetic data, detailed blood chemistry, and dietary information, to predict healthy aging in individuals.

Technical Abstract: Background: Accurate and consistent measurements of what humans eat are essential for health and nutrition research, but typical form- or interview-based measurements often are inaccurate, incomplete and subjective. Although many biomarkers of food intake (BFIs) are validated indicators of intake of specific foods, how food-BFI relationships are affected by genetic and lifestyle factors is insufficiently unexplored. Methods: Dietary intake, clinical, anthropometric, blood metabolomics and genotype data from 782 self-described Puerto Ricans were available. Thirty-one BFI-food intake relationships were assessed with linear regression models, using covariates based on significant covariate-BFI associations. Metabolite-genotype associations for 20 BFIs were identified at a strict genome-wide threshold of 1.00E-08 from 712,197 autosomal single nucleotide polymorphisms. Results: Using epidemiological data, we identified 12 known BFI-food pairs that reached statistical significance for blood biomarkers. Importantly, we adjusted tests for the BFI-food intake relationship with factors demonstrated to associate significantly with the BFI metabolite, including age, body weight, physical activity and sex. For genome-wide association tests of blood metabolites, at a rigorous genome-wide significance threshold, 10 genetic variants were identified to show association with BFIs 3-methylxanthine, gluconic acid, isoleucine and tartaric acid. The most significant associations for each BFI exhibited P values ranging from 3.20E-08 in the KIF11 gene for 3-methylxanthine to 2.55E-15 in CCDC7 for tartaric acid. Conclusion: In a Caribbean Hispanic population, 12 validated BFI-food intake pairs showed statistical significance that were modulated by non-dietary factors. The identification of genotype-BFI associations implies that the applicability of a BFI could depend on common genetic differences. These results provide useful information regarding the incorporation of metabolomics data into precision nutrition models of aging and future health outcomes, particularly for metabolites that function as biomarkers of food intake.