Dr Victor Kunin, Engelbrektson A, Ochman H, Dr Philip Hugenholtz Environ. Microbiol. 10:3p. . August 27, 2009. PMID:19725865 Summary Massively parallel pyrosequencing of the small subunit (16S) ribosomal RNA gene has revealed that the extent of rare microbial populations in several environments, the 'rare biosphere', is orders of magnitude higher than previously thought. One important caveat with this method is that sequencing error could artificially inflate diversity estimates. Although the per-base error of 16S rDNA amplicon pyrosequencing has been shown to be as good as or lower than Sanger sequencing, no direct assessments of pyrosequencing errors on diversity estimates have been reported. Using only Escherichia coli MG1655 as a reference template, we find that 16S rDNA diversity is grossly overestimated unless relatively stringent read quality filtering and low clustering thresholds are applied. In particular, the common practice of removing reads with unresolved bases and anomalous read lengths is insufficient to ensure accurate estimates of microbial diversity. Furthermore, common and reproducible homopolymer length errors can result in relatively abundant spurious phylotypes further confounding data interpretation. Thus, stringent quality-based trimming of 16S pyrotags and clustering thresholds no greater than 97% identity should be used to avoid overestimates of the rare biosphere. You must log in to comment articles |
No rankings yet Recommended by 0 users Download this paper: 2009-wrinkles.pdf RankYou must log in to rank or recommend articles
|