SVCROWS: A User-Defined Tool for Interpreting Significant Structural Variants in Heterogeneous Datasets.

Structural variants (SVs) are abundant across all life, and have major impacts on the genome and transcriptome. However, it is difficult to appreciate the individual significance of SVs when they are heterogeneously distributed across a genomic neighborhood. Further, low-input sequencing technologies or sequencing of many individuals across a population introduce variance that complicates SV counting and association studies. Tools exist to simplify SV datasets, but these SV mergers begin to fail on large or highly variable datasets. To address this issue, we introduce a new SV merger called SVCROWS (Structural Variation Consensus with Reciprocal Overlap and Weighted Sizes). This option-rich R package merges and summarizes SV regions using a size-weighted reciprocal overlap framework, effectively accounting for skewed impacts of variable-length SVs. User input directs stringency of comparisons across a range of sizes, enabling different levels of resolution in complex genome regions that harbor both small and large SVs. When compared to other SV merging programs, SVCROWS accurately merges SVs while maintaining less frequent genotypes of the unmerged SV calls. SVCROWS proves to be especially useful with large and highly variable single-cell datasets for enabling SV discovery. Overall, the novel size-weighted comparisons of SVCROWS presents a framework for improved interpretation of SV calls, and its ease of use allows it to be applied to virtually any upstream analyses.