The Phusion Assembler

Abstract

The Phusion assembler has assembled the mouse genome from the whole-genome shotgun (WGS) dataset collected by the Mouse Genome Sequencing Consortium, at ∼7.5× sequence coverage, producing a high-quality draft assembly 2.6 gigabases in size, of which 90% of these bases are in 479 scaffolds. For the mouse genome, which is a large and repeat-rich genome, the input dataset was designed to include a high proportion of paired end sequences of various size selected inserts, from 2–200 kbp lengths, into various host vector templates. Phusion uses sequence data, called reads, and information about reads that share common templates, called read pairs, to drive the assembly of this large genome to highly accurate results. The preassembly stage, which clusters the reads into sensible groups, is a key element of the entire assembler, because it permits a simple approach to parallelization of the assembly stage, as each cluster can be treated independent of the others. In addition to the application of Phusion to the mouse genome, we will also present results from the WGS assembly ofCaenorhabditis briggsae sequenced to about 11× coverage. TheC. briggsae assembly was accessioned through EMBL,http://www.ebi.ac.uk/services/index.html, using the series CAAC01000001–CAAC01000578, however, the Phusion mouse assembly described here was not accessioned. The mouse data was generated by the Mouse Genome Sequencing Consortium. The C. briggsae sequence was generated at The Wellcome Trust Sanger Institute and the Genome Sequencing Center, Washington University School of Medicine.