Pooled long-read sequencing for structural variant characterization in schistosome populations

Genetic diversity circulates in populations in multiple forms; from single nucleotide variants to genome duplications. Population genetic studies frequently emphasize single nucleotide variants, small indels (<=50bp) occasionally, but typically neglect large structural variants (>50bp), despite the fact that structural variants impact disproportionately large portions of the genome and may have a large impact on phenotype. Here, we examined large structural variants circulating in five lab populations of the human parasite Schistosoma mansoni by generating long-read sequences from pools of worms. Using this method, we were able identify and genotype 17,446 SVs, representing 6.5% of the genome despite challenges in identifying low frequency variants. More than half (59%) of the SVs were shared between >=4 populations, but 12% were found in only one of the five populations. Within this subset, we identified 168 population-specific SVs that were at-or-near fixation in one population but missing (<5%) in the other four populations, and 5 of these variants directly impacted the coding sequence of 6 genes. We also identified 8 structural variants with extreme allele frequency differences between populations in genome regions associated with pathogen phenotypes. These results demonstrate that long-read sequence data from pooled individuals is a viable method to quickly catalogue structural variants circulating within populations and that some of these variants may be responsible for, or linked to regions experiencing, population-specific directional selection.