The extent to which different sequence-based approaches describe environmental microbial communities in comparative studies is an important consideration when deriving inferences from ecological studies. The ability of a targeted metagenomic approach [small subunit (SSU) rRNA pyrosequencing] and shotgun metagenome approaches were compared to identify distinguishing features in dryland soil microbial communities from two different habitats: biological soil crusts (biocrusts) and creosote bush root zones. A parallel comparison was conducted to determine the ability of each approach to detect community differences potentially arising from a more subtle experimental treatment, long-term elevated atmospheric carbon dioxide. As expected, the biocrust datasets were clearly differentiated from root zone datasets using either of the sequencing approaches. However, the composition described by each approach was significantly different. The magnitude of comparative differences due to habitat or elevated CO2 treatment was larger with pyrosequenced SSU datasets or SSU reads recruited from shotgun metagenomes, than from SEED-classified shotgun metagenome reads. Finally, based on prior knowledge of the biocrust communities, the SSU-based datasets more accurately identified the dominant biocrust cyanobacteria sequences compared to the shotgun metagenome datasets.
Published 2012. This article is a U.S. Government work and is in the public domain in the USA.