COS 35-6 - Freshwater bacteria populations respond to seasonal and decadal environmental change over 20 years

Abstract: Long-term time series are required to distinguish decadal changes from short-term seasonal cycles. However, multi-decadal microbial time series, where sample collection began before the advent of high throughput sequencing, are rare. Here, we obtained deep metagenomic sequencing from a 20-year, 471-sample time series from Lake Mendota (WI, USA). This constitutes the longest freshwater metagenome time series to date. To identify genomic changes over different time scales, we obtained over 30,000 medium to high quality ( >50% complete) genomes. These were de-replicated to produce 2,900 representative genomes reconstructed from a broad diversity of bacteria including the dominant lineages in the lake. We mapped reads to these genomes to identify single nucleotide variants (SNVs) within individual samples and characterize genomic microdiversity over time. Lake Mendota’s microbial community follows strong phenological patterns, and abundance estimates based on coverage confirmed this. In addition, we observed seasonal patterns in genomic microdiversity, which were often not concordant with abundance. For example, one strain of the most abundant bacteria in the lake, the Actinobacteria Nanopelagicus (acI-B), consistently peaked in microdiversity in summer, after a spring abundance peak. On top of this genomic phenology, we observed decadal shifts in microdiversity. For example, in another acI-B strain, microdiversity shifted higher at the midpoint of the time series. This shift coincides with two aquatic species invasions that disrupted the lake’s food web, the zooplankton spiny water flea and zebra mussels. This work is a detailed examination of the long-term interplay between microbial ecology and evolutionary processes in nature.