The DRAGEN Germline V2+ Pipeline provides ultra-rapid analysis of NGS data, and reduces the time required to analyze a whole human genome from ~20 hours (BWA-GATK) to ~24 minutes onsite and ~34 minutes in the cloud (AWS F1.16xlarge instance), while achieving industry leading variant calling accuracy. New features in the DRAGEN Germline V2+ Pipeline include improved mapper and alignment algorithms, Alt-Aware mapping for Graph-based references, and a sample-specific error modeling algorithm that greatly improves variant calling accuracy for INDELs.
DRAGEN Germline V2+ Pipeline
The DRAGEN Germline V2+ Pipeline provides ultra-rapid genome sequence analysis with industry-leading accuracy. Aside from algorithmic improvements, the DRAGEN V2+ DRAGEN Germline V2+ Pipeline improves on the DRAGEN Germline V2 Pipeline, which was the top performer in 5 out of 6 categories amongst entrants that identified all 50 hidden variants in the 2017 Precision FDA Hidden Treasures Challenge Warm-Up.
DRAGEN Germline V2+ Pipeline Algorithmic Improvements
The DRAGEN Germline V2+ Pipeline incorporates enhancements in the algorithm that reduce the false positive rate, produce a better f-measure, and overall greater sensitivity. DRAGEN Germline V2+ reduces false positives without the need to eliminate a large number of reads, leading to its ability to generate higher sensitivity. DRAGEN Germline V2+ achieves false positive reduction through:
- The identification of contaminant reads and preventing them from contributing to the variant analysis.
- Identification of mis-mapped reads through patterns of reads that are not consistent with properly mapped reads.
- Detecting regions of the genome where the sequencing technology encounters problems during the base incorporation process.
Other algorithmic improvements include a sample-specific error model that improves variant calling accuracy by improving INDEL accuracy for datasets that have undergone polymerase chain reaction (PCR).
Effects on Sample Coverage: DRAGEN Germline V2+ Pipeline Accuracy
Variant Calling Accuracy: DRAGEN Germline V2+ vs. 3rd Party Analysis Tools
Comparison of SNP Accuracy*
A SNP (single nucleotide polymorphism) is a single base that differs between two genomes. In this comparison using HG001 of DRAGEN Germline V2+ against other third party analysis tools on the market, DRAGEN V2+ demonstrated less SNP errors compared to GATK, Strelka2, and Freebayes.
Comparison of INDEL Accuracy*
An INDEL is the insertion or deletion of one or more base pairs in a sample genome compared to a reference. In this comparison using HG001 of DRAGEN Germline V2+ against other third party tools on the market, DRAGN V2+ demonstrated less INDEL errors than GATK, DeepVariant, and Freebayes.
*Both of these comparisons are independent benchmarks from a trusted third party (DNAnexus) and can be found on the DNAnexus Blog, “How to Train Your DRAGEN – Evaluating and Improving Edico Genome’s Rapid WGS Tools.” March 2018. https://blog.dnanexus.com/2018-03-08-how-to-train-your-dragen-evaluating-and-improving-edico-genomes-rapid-wgs-tools/