Stefan Neuhaus

rRNA and rDNA based assessment of sea ice protist biodiversity from the central Arctic Ocean

Abstract Sea ice is a large and diverse ecosystem contributing significantly to primary production in ice-covered regions. In the Arctic Ocean, sea ice consists of mixed multi-year ice (MYI), often several metres thick, and thinner first-year ice (FYI). Current global warming is most severe in Arctic regions; as a consequence, summer sea ice cover is decreasing and MYI is disappearing at an alarming rate. Despite its apparent hostility, sea ice is inhabited by a diverse microbial community of bacteria and protists, many of which are photosynthetic. Here we present an assessment of eukaryotic biodiversity in MYI and FYI from the central Arctic Ocean using high-throughput 454 sequencing of 18S rRNA and rDNA amplicons. We compared the rDNA-based ‘total’ biodiversity with the ‘active’ biodiversity from rRNA amplicons and found differences between them including an over-representation of Ciliophora, Bicosoecida and Bacillariophyceae operational taxonomic units (OTUs) in the active part of the community. Differences between the two libraries are more pronounced at the lower taxonomic level: certain genera, such as Melosira, are more abundant in the rRNA library, indicating activity of these genera. Furthermore, we found that one FYI station showed a higher activity of potential grazers which was probably due to the advanced stage of melt evident by higher ice temperatures and highly porous ice compared with the other stations.

Impact of sequence processing and taxonomic classification approaches on eukaryotic community structure from environmental samples with emphasis on diatoms

Next‐generation sequencing is a common method for analysing microbial community diversity and composition. Configuring an appropriate sequence processing strategy within the variety of tools and methods is a nontrivial task and can considerably influence the resulting community characteristics. We analysed the V4 region of 18S rRNA gene sequences of marine samples by 454‐pyrosequencing. Along this process, we generated several data sets with QIIME, mothur, and a custom‐made pipeline based on DNAStar and the phylogenetic tree‐based PhyloAssigner. For all processing strategies, default parameter settings and punctual variations were used. Our results revealed strong differences in total number of operational taxonomic units (OTUs), indicating that sequence preprocessing and clustering had a major impact on protist diversity estimates. However, diversity estimates of the abundant biosphere (abundance of ≥1%) were reproducible for all conducted processing pipeline versions. A qualitative comparison of diatom genera emphasized strong differences between the pipelines in which phylogenetic placement of sequences came closest to light microscopy‐based diatom identification. We conclude that diversity studies using different sequence processing strategies are comparable if the focus is on higher taxonomic levels, and if abundance thresholds are used to filter out OTUs of the rare biosphere.