A de novo assembly method for metagenomic DNA reads with mate pairs

摘要

A high quality assembly of reads, which are short fragments generated from shotgun sequencing, is a substantial part of the metagenome projects. Although traditional assemblers have been employed in initial analysis of the metagenomes, they cannot surmount the challenges created by the features of metagenomic data. We present a de novo assembly algorithm for metagenomes with mate pairs. Our method uses mate pair information to guide contig construction. More specifically, we developed an efficient algorithm to find the best mate pair threading paths from the overlap graph to construct contigs. Tests on simulated metagenome data show that our method produced more accurate assembly than Celera Assembler and Phrap when assembling contigs at a same level.