Identification of Variant Compositions in Related Strains Without Reference

Current DNA sequencing technologies do not read an entire chromosome from end to end but instead produce sets of short reads, i.e. fragments of the genome. Haplotype assembly is the problem of assigning each read to the correct chromosome in the set of chromosomes in a homologous group, with the aid of the reference sequence. In this paper, we extend an existing exact algorithm for haplotype assembly of diploid species (Patterson et al, 2014) to the reference-free, polyploid case. A reference-free method does not exploit a reference genomic sequence of a species and thus we cannot exploit a known linear order for the reads and resulting variant positions. Therefore we obtain an unordered variant composition as a result. This setting can be also applied to the study of relative abundances of related bacterial strains.