SRP190199 PRJNA530071 Lake Baikal summer amplicons This study is the first to identify distinct bacterial assemblages in Lake Baikal during summer stratification and link them to specific environments. Multiple linear regression and model averaging show that community diversity is driven by richness in the mixed layer (ML) and evenness in the deep waters (DW), and we identify environmental covariates that explain these contrasting trends. Network analyses reveal assemblages specific to ML and DW, where phylogeny reflects preferred environments. We then use PICRUSt to predict metagenomes and reveal that redundancy is lowest in the coolest areas of ML, potentially placing them at greatest risk of microbial functional diversity loss. This is a significant step towards understanding and addressing challenges faced by microbial communities in a changing world.