Distantly sampled soils carry few species in common.

The bacterial phylogenetic structure of soils from four distinctly different sites in South and North America was analyzed. One hundred and thirty-nine thousand sequences of the V9 region of the small subunit of the bacterial ribosomal RNA gene generated for a previous study were used for this work. Whereas the previous work estimated levels of species richness, this study details the degree of bacterial community overlap between the four soils. Sequences from the four soils were classified and grouped into different phyla and then assigned to operational taxonomic units (OTUs) as defined by 97 or 100% sequence similarity. Pairwise Jaccard and theta similarity indices averaged over all phyla equalled 6 and 12% respectively at the 97% similarity level, and 15% for both at the 100% similarity level. At 100 and 97% sequence similarity, 1.5 and 4.1% of OTUs were found in all four soils respectively, and 87.9 and 74.4%, respectively were a unique particular soil. These analyses, based on the largest soil bacterial sequence retrieval to date, establish the high degree of community structure difference for randomly sampled dissimilar soils and support the idea that wide sampling is important for bioprospecting. The 10 most abundant cultured genera were determined in each soil. These 10 genera comprised a significant proportion of the reads obtained from each soil (31.3-37.4%). Chitinophaga was the most abundant or the second most abundant genus in all four soils with 7.5-13.8% of the total bacterial sequences in these soils. The striking result is that several culturable genera, whose roles in soil are virtually unknown, were found among these dominant sequences.