A Workflow for the Analysis of Contigs from the Metagenomic Shotgun Assembly of SMRT® Sequencing Data

The throughput of SMRT® Sequencing and long reads allows microbial communities to be analyzed using a shotgun sequencing approach. Key to leveraging this data is the ability to cluster sequences belonging to the same member of a community.  Long reads of up to 40 kb provide a unique capability in identifying those relationships, and pave the way towards finished assemblies of community members. Long reads are highly valuable when samples are more complex  and containing lower intra-species variation, such as a larger number of closely related species, or high intra-species variation.

Here, we present a collection of tools tailored for the analysis of PacBio® metagenomic assemblies. These tools allow for improvements in the assembly results, and greater insight into the complexity of the study communities.

Supervised classification is applied to a large set of sequence characteristics (e.g. %GC, read coverage, kmers ) and to cluster contigs from single or highly related species. Assembly in isolation of the raw data associated with these contigs is shown to improve assembly statistics. A unique feature of SMRT Sequencing is the availability to leverage simultaneously collected base modification / methylation data to aid the  clustering of contigs expected to comprise a single or very closely related species.  We demonstrate the added value of base modification information to distinguish and study variation within metagenomic samples based on differences in the methylated DNA motifs involved in the restriction modification system.

Application of these techniques is demonstrated on a mock community and monkey intestinal microbiome sample.