Diversity and phylogenetic structure of two complex marine microbial communities

Abstract

(cont.) competitive mechanisms are too weak to purge diversity from within them.

Molecular surveys have revealed that microbial communities are extraordinarily diverse. Yet, two important questions remain unanswered: how many bacterial types co-exist, and do such types form phylogenetically discrete units of potential ecological relevance? This thesis explores these questions by investigating bacterial diversity in two complex marine communities (coastal bacterioplankton and sediment sulfate-reducing bacteria) by (i) comprehensive analysis of large 16S rRNA clone libraries, and (ii) refinement and application of parametric diversity estimators. Identification and correction of sequence artifacts demonstrated their potentially significant contribution to diversity estimates. Still, hundreds of unique rRNA sequences (ribotypes) were detected in the corrected libraries, and extrapolation to community diversity with commonly used non-parametric diversity estimators suggested at least thousands of co-existing ribotypes in the two communities. However, close inspection revealed that the non-parametric estimators likely lead to underestimation of ribotype diversity in the clone libraries. Thus, an improved parametric method was developed and shown to closely fit the data. The extrapolated total ribotype diversity in the sample by the improved method was up to one order of magnitude higher than estimated with common non-parametric approaches. Most significantly, the compensation for artifacts and improved estimation revealed that the vast majority of ribotypes fall into microdiverse clusters containing <1% sequence divergence. It is proposed that the observed microdiverse clusters form important units of differentiation in microbial communities. They are hypothesized to arise by selective sweeps and contain high diversity because