Fig. S1. Analysis of the 16S rRNA gene controls. A set of three full-length 16S rRNA gene fragments served as controls to test for errors during PCR amplification and pyrosequencing (Garcia et al. 2011). Control DNAs sent to the sequencing company consisted of: (1) Sulfurospirillum multivorans, (2) an actinobacteria strain cultured in the laboratory from freshwater named CD11, (3) DNA from SAG acI-B1 (Garcia et al., 2013) and (4) a mixture of all the three. A total of 10 538 sequences were obtained for the control 16S rRNA gene fragments. The length of the sequences had a median of 228 bp. In an attempt to reduce the number of overestimated OTUs and find the optimal distance at which OTUs should be clustered, five distances were tested. At a distance 0.04, 99.9% of the control sequences were correctly classified, and in total 6 OTUs were observed, still yielding 3 more OTUs than expected from the three different 16S rRNA genes. Increasing the distance to 0.05 led to only very small gains in the number of correctly identified sequences. For this reason, subsequent analyses were performed at distance 0.04 (or 96% identity).


Fig. S2. Top OTUs per month (September, October and November respectively). See Table 1 for further characterization of the OTUs.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.