The assembly of metagenomes is dramatically improved by the long read lengths of SMRT Sequencing. This is demonstrated in an experimental design to sequence a mock community from the Human Microbiome Project, and assemble the data using the hierarchical genome assembly process (HGAP) at Pacific Biosciences. Results of this analysis are promising, and display much improved contiguity in the assembly of the mock community as compared to publicly available short-read data sets and assemblies. Additionally, the use of base modification information to make further associations between contigs provides additional data to improve assemblies, and to distinguish between members within a microbial community. The epigenetic approach is a novel validation method unique to SMRT Sequencing. In addition to whole-genome shotgun sequencing, SMRT Sequencing also offers improved classification resolution and reliability of metagenomic and microbiome samples by the full-length sequencing of 16S rRNA (~1500 bases long). Microbial communities can be detected at the species level in some cases, rather than being limited to the genus taxonomic classification as constrained by short-read technologies. The performance of SMRT Sequencing for these metagenomic samples achieved >99% predicted concordance to reference sequences in cecum, soil, water, and mock control investigations for bacterial 16S. Community samples are estimated to contain from 2.3 and up to 15 times as many species with abundance levels as low as 0.05% compared to the identification of phyla groups.