Large memory high performance computing enables comparison across human gut microbiome of patients with autoimmune diseases and healthy subjects

Abstract

Microbial communities that live on the outside and inside of the human body dramatically influence human health and diseases. In recent years, major progress has been made in understanding the human microbiome communities through projects such as the Human Microbiome Project (http://commonfund.nih.gov/hmp/), using next generation sequencing technologies and metagenomic approaches. In this paper, we describe a comparative computational analysis of 183 human gut microbiome sequence datasets, drawn from healthy individuals as well as those with autoimmune diseases. About 2.4 TB of Illumina deep sequencing metagenomic data were analyzed using computational workflows we developed, which run multiple steps of data- and computing-intensive analyses such as mapping, sequence assembly, gene identification, clustering and functional annotations. The analyses were carried out on the Gordon supercomputer at the San Diego Supercomputer Center (SDSC), using ~180,000 core hours and tens of TB storage space. Our analysis reveals the detailed microbial composition, dynamics, and functional profiles of the samples and provides new insight into how to correlate microbial profiles with human health and disease states.