164
Using the Whole Read: Structural Variant Detection Using NGS Data
Several classes of Structural Variants (SV) remain difficult to detect within sequenced genomes. Deletions and tandem duplications may affect a large proportion of variable genomic sequence space, yet their detection is still difficult to discern from false positive signals. Here, we present a method for detecting such variants from short-read sequence data using the orientation and distance of paired-end, and split-read mappings in addition to using read-depth as a filtering agent. We test our data using simulated SVs and find that our method is 27.5 times more precise than a competing detection program in detecting tandem duplications. Our method is also able to detect three times the number of deletions than a competing algorithm. This high degree of precision should enable better functional prediction of SVs from short-read sequence data.
Keywords:
variants
sequencing
software
