svviz: a read viewer for validating structural variants

Published: August 18, 2015

Author(s)

Noah Spies, Justin M. Zook, Marc L. Salit, Arend Sidow

Abstract

We present svviz, a sequencing read visualizer for structural variants (SVs) that sorts and displays only reads relevant to a candidate SV. This is accomplished by searching input bam(s) for potentially relevant reads, realigning them against the inferred sequence of the putative variant allele as well as the reference allele, and identifying reads that match one allele better than the other. Reads are assigned to the proper allele based on alignment score, read pair orientation and insert size. Separate views of the two alleles are then displayed in a scrollable web browser view, enabling a more intuitive visualization of each allele, compared to the single reference genome-based view common to most current read browsers. The web view facilitates examining the evidence for or against a putative variant, estimating zygosity, visualizing affected genomic annotations, and manual refinement of breakpoints. An optional command-line-only interface allows summary statistics and graphics to be exported directly to standard graphics file formats. svviz is open source and freely available from github, and requires as input only structural variant coordinates (called using any other software package), reads in bam format, and a reference genome. Reads from any high-throughput sequencing platform are supported, including Illumina short-read, mate-pair, synthetic long-read (assembled), Pacific Biosciences, and Oxford Nanopore. svviz is open source and freely available from https://github.com/svviz/svviz.
Citation: Bioinformatics
Volume: 31
Pub Type: Journals
Created August 18, 2015, Updated November 10, 2018