Detection of structural DNA variation from next generation sequencing data: a review of informatic approaches

Next generation sequencing (NGS), or massively paralleled sequencing, refers to a collective group of methods in which numerous sequencing reactions take place simultaneously, resulting in enormous amounts of sequencing data for a small fraction of the cost of Sanger sequencing. Typically short (50-250 bp), NGS reads are first mapped to a reference genome, and then variants are called from the mapped data. While most NGS applications focus on the detection of single nucleotide variants (SNVs) or small insertions/deletions (indels), structural variation, including translocations, larger indels, and copy number variation (CNV), can be identified from the same data. Structural variation detection can be performed from whole genome NGS data or "targeted" data including exomes or gene panels. However, while targeted sequencing greatly increases sequencing coverage or depth of particular genes, it may introduce biases in the data that require specialized informatic analyses. In the past several years, there have been considerable advances in methods used to detect structural variation, and a full range of variants from SNVs to balanced translocations to CNV can now be detected with reasonable sensitivity from either whole genome or targeted NGS data. Such methods are being rapidly applied to clinical testing where they can supplement or in some cases replace conventional fluorescence in situ hybridization or array-based testing. Here we review some of the informatics approaches used to detect structural variation from NGS data.