Improved detection of changes in species richness in high diversity microbial communities.

Biodiversity is important for balance and function of a broad variety of ecosystems, and identifying factors that influence biodiversity can assist environmental management and maintenance. However, low abundance taxa are often missing from ecosystem surveys. These rare taxa, which may be critical to the ecosystem function, are not accounted for in existing methods for detecting changes in species richness. We introduce a model for total (observed and unobserved) biodiversity that explicitly accounts for these rare taxa. Our method permits rigorous testing for both heterogeneity and biodiversity changes, and simultaneously improves type I and II error rates compared with existing methods. To estimate model parameters we utilize the well-developed literature of meta-analysis. The problem of substantial low abundance taxa missing from samples is especially pronounced in microbiomes, which are the focus of our case-studies. [ABSTRACT FROM AUTHOR]