The prevalence of species and strains in the human microbiome: a resource for experimental efforts.

PLoS One. 2014 May 14;9(5):e97279. doi: 10.1371/journal.pone.0097279. eCollection 2014.


Experimental efforts to characterize the human microbiota often use bacterial strains that were chosen for historical rather than biological reasons. Here, we report an analysis of 380 whole-genome shotgun samples from 100 subjects from the NIH Human Microbiome Project. By mapping their reads to 1,751 reference genome sequences and analyzing the resulting relative strain abundance in each sample we present metrics and visualizations that can help identify strains of interest for experimentalists. We also show that approximately 14 strains of 10 species account for 80% of the mapped reads from a typical stool sample, indicating that the function of a community may not be irreducibly complex. Some of these strains account for >20% of the sequence reads in a subset of samples but are absent in others, a dichotomy that could underlie biological differences among subjects. These data should serve as an important strain selection resource for the community of researchers who take experimental approaches to studying the human microbiota


Kraal L, Abubucker S, Kota K, Fischbach MA, Mitreva M.