Introducing Otushuff and DwOdum: A New Set of Tools for Estimating Beta Diversity for Under-Sampled Communities.

Characterization of complex microbial communities by DNA sequencing has become a standard technique in microbial ecology. Yet, particular features of this approach render traditional methods of community comparison problematic. In particular, a very low proportion of community members are typically sampled, and spurious taxa can be generated in datasets through a number of mechanisms (e.g., sequencing errors, incomplete sampling). A robust measure of distance between two microbial communities, as assessed by sequencing surveys, should correct for random differences due to incomplete sampling. We present a new analytical method, OTUshuff, which only considers OTU/taxon differences between communities that are statistically significant. We also propose a new distance measure, the weighted Odum (D_wOdum) distance score. D_wOdum is particularly suited to sequence-based characterization of microbial communities, as it has reduced sensitivity to differences in the abundance of rare taxa. We illustrate the utility of OTUshuff, D_wOdum, and their combination, using simulated data as well as an actual soil microbial community pyrosequencing dataset derived from diverse sites and with plot-level replicates. The D_{OTUshuff-wOdum} approach to estimating distance between communities successfully corrects over-estimation of distances when sampling is incomplete. D_{OTUshuff-wOdum} also removes artificial variation among replicate sub-samples drawn from a common pool of sequence reads, while maintaining the ability to detect differences among plots and among sites.