A comprehensive evaluation of the sl1p pipeline for 16S rRNA gene sequencing analysis

Fiona J. Whelan, Michael G. Surette

Advances in next-generation sequencing technologies have allowed for detailed, molecular-based studies of microbial communities such as the human gut, soil, and ocean waters. Sequencing of the 16S rRNA gene, specific to prokaryotes, using universal PCR primers has become a common approach to studying the composition of these microbiota. However, the bioinformatic processing of the resulting millions of DNA sequences can be challenging, and a standardized protocol would aid in reproducible analyses. The short-read library 16S rRNA gene sequencing pipeline (sl1p, pronounced "slip") was designed with the purpose of mitigating this lack of reproducibility by combining pre-existing tools into a computational pipeline. This pipeline automates the processing of raw 16S rRNA gene sequencing data to create human-readable tables, graphs, and figures to make the collected data more readily accessible. Data generated from mock communities were compared using eight OTU clustering algorithms, two taxon assignment approaches, and three 16S rRNA gene reference databases. While all of these algorithms and options are available to sl1p users, through testing with human-associated mock communities, AbundantOTU+, the RDP Classifier, and the Greengenes 2011 reference database were chosen as sl1p's defaults based on their ability to best represent the known input communities. sl1p promotes reproducible research by providing a comprehensive log file, and reduces the computational knowledge needed by the user to process next-generation sequencing data. sl1p is freely available at https://bitbucket.org/fwhelan/sl1p .