RESUMO
Direct sequencing of mRNA isolated from environmental samples has been commonly used to analyze the functional activity of ambient communities and, occasionally used for taxonomic identification. Here we assess the viability of using mRNA for investigating the species composition of marine benthic eukaryotes. Total RNA was extracted from sediments sampled close to fish farms and mRNA was sequenced after poly-A enrichment on an Illumina MiSeq sequencer. We investigated the origin of both raw reads and assembled contigs by aligning them to the NCBI non-redundant (nr/nt) nucleotide database. Although sequences were predominantly of eukaryotic origin, the analyses were complicated both by experimental and database artefacts. These issues were addressed by applying filtering procedures that removed the majority of ambiguous sequences from downstream analyses. These processes resulted in a set of 436 high-confidence contigs, the vast majority of which mapped to benthic organisms. Our alignments were dominated by annelids, consistent with burrowing groups found in a parallel morphological analysis. This study shows that it is possible to obtain adequate taxonomic information from the RNA of an eukaryotic community from a limited sample at a moderate cost, demonstrates how both laboratory and in silico artefacts can be overcome through appropriate bioinformatic procedures, and finally highlights some of the drawbacks and caveats of using NCBI as a reference database for such a dataset.