Haplotype and minimum-chimerism consensus determination using short sequence data

Background: Assembling haplotypes given sequence data derived from a single individual is a well studied problem, but only recently has haplotype assembly been considered for population-sampled data. We discuss a software tool called Hapler, which is designed specifically for low-diversity, low-coverage data such as ecological samples derived from natural populations. Because such data may contain error as well as ambiguous haplotype information, we developed methods that increase confidence in these assemblies. Hapler also reconstructs full consensus sequences while minimizing and identifying possible chimeric points. Results: Experiments on simulated data indicate that Hapler is effective at assembling haplotypes from gene-sized alignments of short reads. Further, in our tests Hapler-generated consensus sequences are less chimeric than the alternative consensus approaches of majority vote and viral quasispecies estimation regardless of error rate, read length, or population haplotype bias. Conclusions: The analysis of genetically diverse sequence data is increasingly common, particularly in the field of ecoinformatics where transcriptome sequencing of natural populations is a cost effective alternative to genome sequencing. For such studies, it is important to consider and identify haplotype diversity. Hapler provides robust haplotype information and identifies possible phasing errors in consensus sequences, providing valuable information for population studies and downstream usage of resulting assemblies.