Diet as a driver of natural selection in human evolution: a paleo-empirical perspective

Ancient DNA-based selection scans have revealed multiple diet-associated single nucleotide polymorphisms (SNPs) that exhibit robust signatures of natural selection ^1–8 . To test whether the selection was indeed driven by diet, we conducted a paleo-empirical investigation. We compiled an isotopic dataset comprising 6,064 ancient human samples and 5,635 food resource samples from Britain. We developed a Bayesian mixing model to estimate individual dietary proportions based on isotopic data and constructed a temporal dietary model spanning the last 10,000 years. Using 1,038 ancient DNA samples, we reconstructed derived allele frequency trajectories for 20 strongly selected SNPs via bootstrap resampling ⁹ . We then applied a generalized additive model (GAM) to estimate both mean and time-varying selection coefficients, while accounting for various evolutionary forces beyond natural selection. Finally, we applied the convergent cross mapping (CCM) ¹⁰ algorithm for causal discovery between the time-varying selection coefficients and their corresponding dietary variable. The findings demonstrate that specific dietary elements, including marine resources, C ₃ plants, and dairy consumption, have acted as selective pressures on specific SNPs whose associated phenotypes are primarily shaped by diet. However, diet-driven signals become less detectable when the phenotypes are influenced by multiple environmental factors, including but not limited to diet.