De novo identification of 'heterotigs' towards accurate and in-phase assembly of complex plant genomes

摘要

Accurate and in-phase de novo assembly of highly polymorphic diploid and polyploid plant genomes remains a critical yet unsolved problem. "Out-of-the-box" assemblies on such data can produce numerous small contigs, at lower than expected coverage, which are hypothesized to represent sequences that are not uniformly present on all copies of a homologous set of chromosomes. Such "heterotigs" are not routinely identified in current assembly algorithms and could be used for haplotype phasing and other assembly improvements for such genomes. We introduce an algorithm which attempts to robustly identify heterotigs present in the assembly of a highly polymorphic diploid organism. The algorithm presented is for use with the 454 platform and for diploid assembly, but is readily adaptable to other sequencing platforms and to polyploid assembly.