On Tuesday, March 21st we hosted a webinar regarding our newly released DRAGEN Virtual Long-Read Detection (VLRD) Pipeline, the first pipeline of its kind in analyzing segmental duplications. Segmental duplications are two or more regions of the genome at least 1 kb in length with >90% sequence similarity. These regions are clinically relevant to analyze as variants in these regions can be linked to rare disease. However, segmental duplications present analysis challenges to researchers because these short reads can map to more than one location on the reference. Prior to the DRAGEN VLRD Pipeline, long-read sequence data, as those produced by PacBio or 10X Genomics sequencers, would have been required to eliminate the challenge of mapping to more than one location.
In this webinar, Rami Mehio, our VP of Engineering here at Edico Genome, talked about how the DRAGEN VLRD Pipeline was specifically designed to call variants in segmental duplications. The pipeline works by jointly calling all regions that are similar, even those with low MAPQ scores as seen in segmental duplications due to their similarity, and then solves for the most likely haplotypes in these regions before proceeding to variant calling.
The DRAGEN VLRD Pipeline shows greater accuracy than traditional variant callers and has gone head-to-head with GATK long-read sequence data. The DRAGEN VLRD Pipeline has demonstrated the ability to achieve greater than or equal accuracy analyzing shorter read lengths (50 bp and 100 bp) than GATK at longer read lengths (200 bp, 400 bp, 600 bp, 800 bp).
You can view the webinar and see the results from the analysis here.