DRAGEN CS includes second-generation versions of the DRAGEN Germline and Somatic Pipelines, which feature enhanced mapping and aligning algorithms and greatly improved variant calling. The new variant calling algorithms are able to better distinguish real variants from errors introduced in sample preparation and sequencing. The DRAGEN Germline V2 Pipeline was recognized as a top performer in the PrecisionFDA Hidden Treasures – Warm Up Challenge in October 2017, receiving the top score in five of six accuracy metrics among 30 participants.
“By utilizing sample-specific prep and sequencer-error modeling, DRAGEN is able to identify real variants with even greater precision, resulting in enhanced accuracy and accelerated analysis speeds,” said Mr. Mehio. “In fact, DRAGEN’s INDEL calling, which usually presents a greater challenge for analysis tools than single nucleotide polymorphisms, demonstrated the highest level of accuracy among pipelines identifying hidden variants in the recent PrecisionFDA challenge. With the integration of these cutting-edge algorithms, DRAGEN is now the top performer in accuracy as well as speed.”
DRAGEN CS contains tools for all pipeline steps, including mapping/aligning, position sorting, duplicate marking and variant calling. The application accepts sequencing data inputs in BCL, FASTQ, and BAM/CRAM formats, and features BCL conversion, download and upload streaming, and compressed hash tables for a more streamlined and efficient workflow. The Complete Suite offers the following pipelines:
- DRAGEN Germline V2 Pipeline provides clinical-grade, end-to-end (BCL to VCF) analysis of whole genome, exome and targeted panel NGS data.
- DRAGEN Somatic V2 Pipeline includes tumor-only and tumor/normal modes, designed for use in detecting somatic variants in tumor samples.
- DRAGEN RNA Gene Fusion Detection Pipeline performs transcriptome analysis starting with splice junction discovery and alignment, followed by gene fusion detection.
- DRAGEN Population Calling Pipeline calls variants jointly across multiple genomes and can scale to thousands of samples at expedited speeds.
- DRAGEN Virtual Long Read Detection (beta release) is an innovative variant caller specialized in mutation detection in segmental duplication regions, such as pseudo-genes. Its accuracy in such regions is equivalent to traditional variant callers, using reads that are 6 to 8 times longer.
- GATK Best Practices Pipeline allows users to run the GATK pipeline at far greater speeds than possible on central processing units (CPUs).
- MuTect2 Pipeline is an enhanced version of MuTec2, a somatic SNP and indel caller.
“Edico Genome’s goal is to simplify NGS for our users – arming them with industry leading speeds, accuracy and scalability through an easy-to-use interface to streamline their work,” said Dr. van Rooyen. “Our engineering team is perpetually working on new algorithms that rapidly and accurately analyze genomic data — enabling clinicians to get to diagnoses faster, and researchers to get to results more rapidly — and we look forward to releasing new offerings in the coming months.”
DRAGEN leverages field programmable gate arrays (FPGAs) to rapidly accelerate secondary analysis of NGS data both onsite and in the Cloud via AWS Marketplace. Edico Genome recently demonstrated the speed and scalability of DRAGEN on the AWS Marketplace by deploying 1,000 Amazon EC2 F1 instances to analyze 1,000 whole human genome sequences in only 2 hours and 25 minutes, setting the GUINNESS WORLD RECORDSTM title for Fastest time to analyze 1,000 human genomes.
A one-day trial for use of exome and genome pipelines is available to new customers. To learn more about DRAGEN on AWS Marketplace, visit http://edicogenome.com/awsmarketplace/.