A comprehensive census of lake microbial diversity on a global scale

Despite recent interest in microbial diversity and community structure of lakes across various spatial scales, a global biogeographic distribution pattern and its controlling factors have not been fully disclosed. Here, we compiled and analyzed 88,334,735 environmental 16S rRNA sequences from 431 lakes across a wide range of geographical distance and environmental conditions (in particular, salinity, 0-373.3 g L-1). Our results showed that lake sediments inhabit significantly (ANOVA: P<0.001) more diverse microbial communities than lake waters. Non-metric dimensional scaling (NMDS) ordinations indicated that microbial community compositions differed distinctly among sample types (freshwater vs. saline, water vs. sediment) and geographic locations. Mantel and partial Mantel tests showed that microbial community composition in lake water was significantly (P=0.001) correlated with geographic distance, salinity, and pH. Statistical analyses based on neutral community and null models indicated that stochastic processes may play predominant roles in shaping the microbial biogeographic distribution patterns in the studied global lake waters. The dispersal-related stochasticity (e.g., homogenizing dispersal) exhibited a stronger influence on the distribution of microbial community in freshwater lakes than in saline lakes. Overall, this work expands our understanding of the impact of geographic distance, environmental conditions, and stochastic processes on microbial distribution in global lakes.