Back

Diet-responsive proteogenomic effects following short-term restriction of animal products in humans

Simistiras, A.; Bocher, O.; Emmanouil, C.; Skoulakis, A.; Glentis, S.; Scarmeas, N.; Zeggini, E.; Rouskas, K.; Dimas, A. S.

2025-07-25 genomics
10.1101/2025.07.21.665884 bioRxiv
Show abstract

The effect of diet on genetic regulation in humans remains largely unexplored. Here, we investigated gene-by-diet interactions in a unique group of apparently healthy individuals (N=200) who alternate between omnivory and dietary restriction of animal products for religious reasons. Using longitudinal plasma proteomic and genotypic data, we identified diet-responsive cis-protein quantitative trait loci (cis-pQTLs) including gain of genetic regulatory effects for LBR and MSRA, proteins linked to cholesterol and methionine metabolism respectively. LBR-associated diet-responsive cis-pQTL rs74148404 showed suggestive evidence of colocalization with obesity exclusively under dietary restriction, suggesting diet-dependent differential genetic risk for disease. Additionally, a novel dietary restriction-associated cis-pQTL for metabolic regulator FGF21 colocalized with eosinophil and platelet traits pointing to diet-sensitive immunometabolic signalling. By parallel profiling of a continuously omnivorous control group (N=211), we also uncovered seasonally dynamic gain and loss of genetic regulation for proteins linked to apoptosis in immune system pathways (MAVS, CASP3, PDLIM7 IL12RB1), effects likely masked by animal product restriction. These findings reveal dynamic diet- and season-sensitive regulatory mechanisms with implications for precision nutrition and individualized disease prevention strategies, and underscore the need to integrate environmental context into genetic studies of health and disease.

Published in Nature Communications (predicted rank #1) · training set

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.