Development of a Computational Method for Identification of Genomic Structural Variants Using Next Generation Sequencing

Abstract

Structural variants (SVs) are changes in genome like deletions, insertions, inversions, duplications and translocations. These changes are in part of genome greater than 1kb of size. SVs that are rearrangements in the genome are responsible for various disorders. SVs not only affect genes but also change the expression of the genes and cana lead to diseases like autism, schizophrenia, rheumatoid arthritis and cancer. To determine these variants Next Generation Sequencing (NGS) is the most promising approach. To identify SVs commonly four different approaches; read pair, read depth, split read and assembly are used. Read pair approach is based on evaluating the distance and orientation of read pairs it uses discordant read pairs to identify SVs. Read depth approach make use of some random distribution in mapping depth to detect variants. Split read approach use single base pair resolution to determine SVs. Assembly based approach work by first reconstructing DNA fragments contigs and then assembles to reference genome to find discordant pairs. Different tools are available to identify SVs these tools use only one or two approaches to identify SVs. The evaluation of most of the tools suggests that the tools that use multiple approaches can identify the variants more accurately. Therefore, SV identifier (SVI), a new computational method based on multiple approaches, to identify SVs is designed using NGS data. SVI is a python based method. This method is designed to detect the SVs like deletions, insertions, inversions and duplications from paired end NGS data. It can analyze multiple samples, it is not platform dependent, detect deletions in reference genome, annotate the variants and also provide visualization of results.