Estimation of viral population structure from amplicon-based reads

Accurately estimating the structure of highly diverse viral populations is a challenging task. There are two main impediments to globally reconstructing a population. The first is the presence of sequencing errors in reads. Judiciously differentiating these errors from actual rare variants must be properly handled or the global structure may be ill-defined. Secondly, long conserved regions in the viral genome extend beyond what modern sequencers are capable of producing. As a result, the actual population diversity may be hidden in these targeted regions. We propose VirA, a tool for global reconstruction of a viral population that overcomes these obstacles by combining local error correction and a read-graph approach.