In this work, we compare the resolution of V2-V3 and V3-V4 16S rRNA regions for the purposes of estimating microbial community diversity using paired-end Illumina MiSeq reads, and show that the fragment, including V2 and V3 regions, has higher resolution for lower-rank taxa (genera and species). It allows for a more precise distance-based clustering of reads into species-level OTUs. Statistically convergent estimates of the diversity of major species (defined as those that together are covered by 95% of reads) can be achieved at the sample sizes of 10000 to 15000 reads. The relative error of the Shannon index estimate for this condition is lower than 4%.

The pelagic zone of Lake Baikal is an ecological niche where phytoplankton bloom causes increasing microbial abundance in spring which plays a key role in carbon turnover in the freshwater lake. Co-occurrence patterns revealed among different microbes can be applied to predict interactions between the microbes and environmental conditions in the ecosystem. We used 454 pyrosequencing of 16S rRNA and 18S rRNA genes to study bacterial and microbial eukaryotic communities and their co-occurrence patterns at the pelagic zone of Lake Baikal during a spring phytoplankton bloom. We found that microbes within one domain mostly correlated positively with each other and are highly interconnected. The highly connected taxa in co-occurrence networks were operational taxonomic units (OTUs) of Actinobacteria, Bacteroidetes, Alphaproteobacteria, and autotrophic and unclassified Eukaryota which might be analogous to microbial keystone taxa. Constrained correspondence analysis revealed the relationships of bacterial and microbial eukaryotic communities with geographical location.

The composition of bacterial communities in Lake Baikal in different hydrological periods and at different depths (down to 1515 m) has been analyzed using pyrosequencing of the 16S rRNA gene V3 variable region. Most of the resulting 34 562 reads of the Bacteria domain have clustered into 1693 operational taxonomic units (OTUs) classified with the phyla Proteobacteria, Actinobacteria, Chloroflexi, Bacteroidetes, Firmicutes, Acidobacteria and Cyanobacteria. It has been found that their composition at the family level and relative contributions to bacterial communities distributed over the water column vary depending on hydrological period. The number of OTUs and the parameters of taxonomic richness (ACE, Chao1 indices) and diversity (Shannon and inverse Simpson index) reach the highest values in water layers. The composition of bacterial communities in these layers remains relatively constant, whereas that in surface layers differs between hydrological seasons. The dynamics of physicochemical conditions over the water column and their relative constancy in deep layers are decisive factors in shaping the pattern of bacterial communities in Lake Baikal.