《CAREER ABI Assembling Extremely》.pdfVIP

CAREER: ABI: Assembling Extremely Large Metagenomes C. Titus Brown Monday, July 23rd, 2012 Keywords: assembly, bioinformatics, next generation sequencing, metagenomics, CAREER: ABI: Assembling Extremely Large Metagenomes PI: C. Titus Brown Lead Institution: Michigan State University Project Summary Complex microbial populations participate in and drive many biochemical and geochemical pro- cesses, ranging from greenhouse gas flux, to nitrogen fixation in soil, to processing of gut nu- trients, and beyond [NRC, 2007]. Only recently has it become possible to deeply sample the contents of these populations using next-generation sequence analysis [Mackelprang et al., 2011, Tyson et al., 2004]. Reference-free assembly of these metagenomes is a critical endeavor in mod- ern biology, in part because we have yet to sample even a small fraction of the tree of life, and have no reference genomes for most environmental organisms. De novo assembly techniques have not kept up with the advances in sequencing. A large class of modern assemblers, de Bruijn graph assemblers, has been developed for the express purpose of short-read assembly and can scale to assemble single human genomes on commodity hardware [Miller et al., 2010, Gnerre et al., 2011]. However, these assemblers are neither designed for nor scale to the volume of data being generated for metagenomes, which can contain many times the novel sequence in genomic samples. Scaling metagenome assembly is an important bioinformatic problem. Research Objectives: I propose a research plan centered on combining a compressible graph representation with novel streaming online data reduction and graph analysis algorithms to provide a general scaling solution to the problem of metagenome assembly. We will combine these novel and existing approaches to develop a usable reference impleme