A comparison of local pipelines and databases for functional metagenomics analyses. Mock Microbiome

One of the promises of shotgun metagenomics is that it will allow assessment of the functional capacities of microbial communities. The analysis of these data, however, is complex, and alternative computational pipelines have been proposed to accomplish this task. All of them make use of homology searches against available, functionally-annotated databases as their main step, but can incorporate other computational steps. In this report we compare the performance of different methods for conducting a functional metagenomics analysis by means of alternative local pipelines and the two large metabolic pathway databases, KEGG and MetaCyc, available, emphasizing their effect on final results. To provide a baseline for the comparison, we have constructed a mock dataset from real, publicly available Illumina sequencing reads. Given that these reads derive from annotated sequence projects, they can be used to validate the metagenomic functional assignment pipelines. Our analysis offers recommendations for optimal functional assignment within metagenomic datasets obtained with the Illumina technology.