Abstract
Nutrition plays an important role in human metabolism and health. However, it is still unclear to what extent self-reported dietary intake is reflected by the metabolite and proteome composition of the human body.
To investigate this question on an epidemiological scale, we generated one of the largest single-site datasets consisting of self-reported dietary habits, absolutely quantified targeted metabolomics (175 metabolites), and relatively quantified proteomics (148 proteins) on serum and plasma samples from 3423 participants of the Cooperative Health Research in South Tyrol (CHRIS) study. Serum metabolite levels represent a snapshot of an individual's current metabolic state. The plasma proteome contains proteins responsible for nutrient transport, responds to lifestyle intervention and immune system activity and is thus attractive for marker assay development in clinical and preclinical research. Using data from a 29-item qualitative food frequency questionnaire, consumption of 29 different foods and food groups, as well as 4 summary food indices (foods grouped by nutrient content) and an overall score for a healthy diet, were analyzed for associationwith measured metabolites and proteins. Principal component analyses and linear regression modeling were used to identify dietary patterns and strongly associated metabolites and proteins, respectively, and will be discussed.
In summary, this study allows consolidation of sets of metabolites and proteins related to participants’ diet and comparison to findings from previous diet and omics related population studies. Metabolomics and proteomics analyses can aid questionnaire-based epidemiological studies by providing related but more objective additional information on participant’s dietary habits.