Back

Candidate Proteins Linking Plant-Based Diet Quality to Cancer, Type 2 Diabetes, and Cardiovascular Disease: Integrating Evidence from Proteomic Analysis and Mendelian Randomization in the EPIC Study

Ler, P.; Matta, K.; Stein, M. J.; Peruchet-Noray, L.; Wu, D.; Gan, Q.; Beigrezaei, S.; Lill, C. M.; Masala, G.; Ricceri, F.; van der Schouw, Y. T.; Verschuren, W. M. M.; Vineis, P.; Jimenez Zabala, A. M.; Zamora-Ros, R.; Tong, T. Y.; Papier, K.; Gunter, M. J.; Viallon, V.; Ferrari, P.; Kim, J.; Freisling, H.

2026-08-11 nutrition
10.64898/2026.08.10.26359939 medRxiv
Show abstract

Plant-based diets may benefit planetary and human health. However, the molecular pathways linking plant-based diet quality to chronic diseases remain unclear. In 4,372 participants from the European Prospective Investigation into Cancer and Nutrition, we assessed 7,285 SomaScan-measured aptamers to identify circulating proteins associated with healthful (hPDI) and unhealthful (uPDI) plant-based dietary patterns using complementary statistical and machine-learning approaches. We then applied cis-pQTL Mendelian randomization (MR) and colocalization to prioritize diet-associated proteins with genetic evidence for associations with overall and site-specific cancers, type 2 diabetes (T2D), and cardiovascular disease (CVD). Among diet-associated proteins with MR evidence, 8 hPDI- and 12 uPDI-related protein-disease associations showed strong colocalization, including EGFR-breast cancer, MMP10-endometrial cancer, NCAN-T2D, and PCSK9-CVD. These findings identify candidate proteins that may link plant-based diet quality to chronic diseases and provide biological insights into the potential health benefits of healthful plant-based diets, which are increasingly relevant to public and planetary health.

Matching journals

The top 6 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.