New DRAGEN v4.4 - Comprehensive genomic analysis with accuracy, speed, and ease-of-use

Author:

Deepthi Shankar

In the last few years Illumina DRAGEN secondary analysis has established incredible gains in accuracy and comprehensiveness. Recent peer reviewed publication in Nature Biotechnology showcased how DRAGEN provides a vastly more comprehensive and accurate genome and variant detection than other solutions available today. DRAGEN v4.4 now powers an even more comprehensive and accurate genome with simplicity and ease of use, providing rich capabilities for emerging multiomic applications and clinical research workflows.

Let’s take a closer look at some of the major updates in DRAGEN v4.4:

Easy-to-use oncology applications

Pre-configured push-button analysis for Heme WGS and Solid WGS tumor-normal - The new DRAGEN Heme somatic tumor-only pipeline (SNV/SV/CNA/Biomarkers) is the first and only commercially available¹ Heme WGS application. This is crucial as WGS provides a comprehensive genomic characterization for hematologic malignancies, well beyond what cytogenetics applications and other methods used today to achieve complete results². With this new release, we simplify the Heme WGS workflow for customers by offering an easy-to-use prepackaged application on the DRAGEN on-premises server (and available on Illumina Connected Analytics) abstracting all the complexities of launching separate pipelines and managing files. The application can be integrated with Connected Insights for a streamlined end-to-end sample to insights workflow.

Figure 1 Out of the box DRAGEN Heme prepackaged application abstracts the complexity of building your own Heme pipeline

Out of the box WGS DNA tumor-normal application (early access) - DRAGEN v4.4 comes with a new pre-packaged application with recommended settings and resource files for WGS solid sample that simplifies WGS tumor-normal workflow on DRAGEN on-premises server, offering both somatic and germline results in a single workflow. It also supports various input file types including FastQ, BAM, CRAM and integrates with Illumina Connected Insights, providing a streamlined sample to insights workflow.

Enhanced multiomic analysis

DRAGEN v4.4 has optimized analysis pipelines to support the new and emerging multiomic assays such as Illumina Single Cell 3’RNA, Illumina Protein Prep, Illumina 5-Base solution and the upcoming spatial transcriptomics assay. An easy-to-use and proven high-quality DRAGEN secondary analysis pipeline optimized for the assay, is included with each assay, streamlining analysis workflow. Here we highlight two analysis pipelines.

DRAGEN Single Cell RNA pipeline available on on-premises server, ICA and BSSH, and included with Illumina single cell 3’RNA prep, enables customers to analyze multiple scRNA preps simultaneously with state-of-the-art methods and visualizations. Hardware acceleration provides efficient secondary analysis of large kits offering a 10-fold decrease in analysis times when compared to the open source PipSeeker solution. The DRAGEN single cell RNA pipeline also provides highly accurate and visual reports for secondary analysis QC, making analysis accessible and easy to use for researchers.

Table 1: Analysis times for DRAGEN vs open source Pipseeker solution

Figure 2 DRAGEN Single Cell RNA pipeline offers visual QC report with information such as trimming, FastQC, Mapping,siingle Cell QC metrics, barcode rank plot, and UMAP plots

DRAGEN Protein Quantification pipeline* included with the Illumina Protein Prep assay is designed to offer accurate and reliable protein detection by including normalization and calibration steps to minimize assay and sample bias. The pipeline is available on ICA/BSSH or on-premises servers** and generates ADAT files with normalized protein counts, which can be further analyzed using Illumina Connected Multiomics.

The DRAGEN protein pipeline outputs two types of files for analyzing Illumina Protein Prep data.

ADAT files from each normalization step
DRAGEN QC report summarizing the QC of the run.

Figure 3- DRAGEN reports highly accurate and visual reports for secondary analysis QC with normalized protein counts

*DRAGEN Protein Quantification pipeline is integrated into Illumina Protein Prep and is currently on DRAGEN v4.3.13 patch

**Available for separate installation

Accuracy Gains

Best DRAGEN structural variant calling accuracy, yet - DRAGEN v4.4 brings a major leap in structural variant calling accuracy with a 30% boost to SV calling accuracy driven by the multigenome mapper with pangenome reference. These advances help close the gap in structural variant calling accuracy between short and long-read sequencing capabilities.

Figure 4 - Accuracy improvements of DRAGEN SV calling; 1. Compared against HG002 NIST T2T Q100 v1.1_v0.019 (hg38) with Truvari v.4.2.2

Personalized pangenome - Personalized pangenome reference is now generally available for all users and provides significant accuracy gains with optimized run times.

Figure 5 - Personalized pangenome and ML Model in DRAGEN 4.4 results in 17.76% reduction in FP+FN (SNPs+INDELs)

DRAGEN v4.4 also offers improved sensitivity and specificity in CNV detection with in-run construction of CNV panel-of-normals (PON) for germline enrichment, offering workflow-level support for automatically constructing a PON from a batch of samples

AWS F2 Instances

Amazon Web Services (AWS) recently launched new AWS EC2 F2 FPGA instance types that deliver faster turnaround and improved scalability with 3rd generation AMD EPYC (Milan) processors. These new F2 instances offer up to 3x as many processor cores, up to twice as much system memory and NVMe (non volatile memory express) storage, and up to 4x the network bandwidth. DRAGEN v4.4 will natively support AWS F2 instances on ICA, BSSH and Bring Your Own License (BYOL) instances on AWS.

Figure 6 DRAGEN on AWS EC2 F2 performance with significant reduction in turn-around time across all workflow types

Other key enhancements in DRAGEN v4.4 include

Streamlined WGS Germline with all targeted callers in standard WGS DRAGEN workflow
New WES targeted calling with support for HBA and SMN
Faster DRAGEN ORA Compression, with upto 30% faster BCL to FASTQ.ORA conversion
Improved accuracy and significantly faster bulk RNA gene fusion detection when compared to DRAGEN v4.3 - 22% reduction in false positives, 9% reduction in false negatives when compared against DRAGEN v4.3, with an average run time of 1 min 20s.
RNA splice variant caller has increased sensitivity when compared to DRAGEN v4.3, enabled with new ML model and mapper updates and more.

Sign up for the webinar to learn about the updates in DRAGEN v4.4

Learn more about DRAGEN here

For Research Use Only. Not for use in diagnostic procedures.

CAP# M-GL-03598

1 As compared to other commercial secondary analysis vendors

2 https://www.illumina.com/company/news-center/feature-articles/WGS-for-heme.html