Navigation
Artificial intelligence driven definition of food preference endotypes in UK Biobank volunteers is associated with distinctive health outcomes and blood based metabolomic and proteomic profiles
A person’s dietary patterns can be influenced by their food choices, which can affect both the hazards and advantages to their health. A Food Preferences Questionnaire (FPQ) was used in this study by Navartilova, Whetton & Geifman (2024) to analyze food preferences using data from more than 180,000 UK Biobank participants. Using Latent Profile Analysis (LPA), the authors identified three distinct dietary profiles: the Omnivore group (strong preference for all foods), the Sweet-tooth group (strong preference for sweet foods and sugary drinks), and the Health-conscious group (strong preference for fruits and vegetables, low preference for sweet or animal-based foods). Blood biochemistry was compared across groups using the non-parametric Kruskal–Wallis test, and the Limma algorithm was employed for differential analysis of 168 metabolites and 2923 proteins. Biological processes and pathways were identified through the DAVID database, and relative risks (RR) for chronic diseases and mental health conditions were calculated, adjusting for sociodemographic factors. The health-conscious group was found to have lower odds of chronic renal disease and heart failure. On the other hand, the sweet-tooth group had higher odds of experiencing depression, stroke, and diabetes. Cancer risk did not differ significantly across the groups. The health-conscious group also had lower inflammatory biomarkers and higher levels of beneficial factors such as ketone bodies and insulin-like growth factor-binding proteins (IGFBP), suggesting that diet may enhance certain metabolic pathways. According to these results, those who choose healthier foods have better health outcomes than those who belong to the Omnivore or Sweet-tooth categories. [NPID: Biomarkers, food preferences, Latent Profile Analysis, metabolomics, proteomics, relative risk, unsupervised machine learning]
Year: 2024
