PMC

Rarefaction curves for the bacterial, fungal, and archaeal clone libraries constructed from each of the soil samples. Rarefaction curves were generated using EstimateS (version 7.5; R. K. Colwell, http://purl.oclc.org/estimates). In all nine libraries, there is no apparent asymptote in the rarefaction curves, suggesting that the libraries do not encompass the full extent of OTU richness in each of the communities with an OTU defined at the ≥97% sequence similarity level.

Predicted values of Simpson's diversity index for each of the 12 communities. Since Simpson's index (D) is defined as the probability that two individuals taken at random from the community belong to the same species (or, in this case, OTU) (), higher values of D⁻¹ indicate higher overall diversity. Symbols correspond to soil type (▴, prairie; ▪, rainforest; •, desert). The mean value for D⁻¹ (with one standard error in parentheses) for each taxonomic group is denoted above each set of symbols.

Hierarchical clustering showing the phylogenetic distance between viral communities from soil (this study), marine sediment (), human fecal samples (), and seawater environments (). Distances were estimated with the weighted Unifrac algorithm (, ) using only those sequences from the metagenomic libraries with significant hits to the Phage Proteomic Tree (http://phage.sdsu.edu/oceanviruses) to generate the input phylogenetic trees. A sequence jackknifing technique was applied to each cluster to determine the sensitivity of the relationships to sample size. Asterisks indicate that the nodes are well supported, having been observed in >95% of the jackknifing runs. The soil viral communities were significantly different from the viral communities in the other environments (P < 0.02 in all cases with the UniFrac significance test) ().

Estimation of OTU richness (M in equation ) (panel a) and the abundance of the most common OTU (a in equation ) (panel b) in each of the three soils. Symbols correspond to soil type (▴, prairie; ▪, rainforest; •, desert). Parameters were estimated by fitting a power law function to OTU abundance distributions. Maximum-likelihood values are denoted with symbols, and bars indicate 68% confidence regions for the parameter estimates of the actual community (see Materials and Methods). Due to the high range of isolikelihood estimates for OTU richness in the desert archaeal, prairie fungal, and rainforest viral communities, we can conclude only that the number of OTUs in each of these communities is likely to exceed 10⁶. The asterisks indicate that the maximum-likelihood estimates of OTU richness for the desert archaeal and rainforest viral communities exceeded 10¹⁰ OTUs.