Slide 1: DRAGEN v4.4 – Introducing our most advanced and comprehensive secondary analysis
DRAGENTM v4.4 – Introducing our most advanced and comprehensive secondary analysis Shyamal Mehtalia Director, Product Management, Secondary Analysis © 2025 Illumina, Inc. All rights reserved. M-GL-03590 For Research Use Only. Not for use in diagnostic procedures.
Slide 2: Agenda
Agenda This webinar will cover: 01 Overview of DRAGEN Secondary Analysis 04 DRAGEN Multiomics 02 DRAGEN Germline 05 DRAGEN Availability on AWS F2 03 DRAGEN Oncology 06 Significant Toolkit and Platform Updates For Research Use Only. Not for use in diagnostic procedures.
Slide 3: Overview of DRAGEN Secondary Analysis
Overview of DRAGEN Secondary Analysis 3 For Research Use Only. Not for use in diagnostic procedures.
Slide 4: A powerful bioinformatics software suite connecting your entire genomics workflow
A powerful bioinformatics software suite connecting your entire genomics workflow PREPARE SEQUENCE ANALYZE INSIGHTS AND REPORTING Clinical Research Connected Oncology Insights Connected Analytics Clarity LIMS BaseSpace Emedgene Genetic Disease Sequence Hub DRAGEN Connected Population Analytics - Gen
Slide 5: DRAGEN secondary analysis
DRAGEN secondary analysis Accurate, comprehensive, efficient NGS secondary analysis Accurate • Winner of 4 Precision FDA challenges1 for germline and somatic pipelines • Highly accurate germline variant calling 99.90%2 F1 score with DRAGEN v4.4 Comprehensive • Tools for genomic, epigenomic, transcri
Slide 6: DRAGEN methods make it a comprehensive analysis choice
DRAGEN methods make it a comprehensive analysis choice ✓ Pangenome Reference ✓ Machine Learning ✓ Targeted Callers ✓ Mosaic calling Read article For Research Use Only. Not for use in diagnostic procedures. 6
Slide 7: DRAGEN is available on different platforms
DRAGEN is available* on different platforms – giving you the flexibility you need On premises server Illumina cloud Public cloud Onboard sequencers Fast, secure, local analysis in a Flexible, secure, scalable Secure cloud analysis with bring managed cloud analysis with NovaSeq X Series, NextSeq frac
Slide 8: DRAGEN – Genetic Disease, Oncology, Multiomics, Toolkit
DRAGEN – Genetic Disease, Oncology, Multiomics, Toolkit Genetic Disease Research Oncology Research Emedgene Connected Insights Trio/Joint Family Heme WGS WES Germline Solid WGS T/O and T/N WGS Germline Minimal Residual Disease (Including Constellation) Multigenome mapper 5 Base (with pangenome refer
Slide 9: The most comprehensive secondary analysis solution
The most comprehensive secondary analysis solution1 General Tools Genetic Disease Oncology Multiomics Graph Capable Mapping Single Nucleotide Variant Single Nucleotide Variant Bulk RNA 2 • Analyze whole genomes, exomes, Adapter Trimming Indel Structural Variant Single-cell RNA panels, transcriptomes
Slide 10: DRAGEN v4.4 with easy-to-use oncology apps, enhanced multiomics, AWS F2 support
DRAGEN v4.4 with easy-to-use oncology apps, enhanced multiomics, AWS F2 support, and more Genetic Diseases Oncology Multiomics & more 30% boost to SV calling accuracy Easy-to-use oncology apps with Support for new multiomic assays with multigenome mapper with push-button analysis for Heme WGS – SIng
Slide 11: DRAGEN – A legacy of continuous innovation and improving accuracy
FP+FN, SNPs+Indels, HG002 NIST v4.2.1 difficult-to-map regions DRAGEN – A legacy of continuous 60,000 52,884 DRAGEN-FN innovation and improving accuracy DRAGEN-FP in each version 50,000 NON-DRAGEN-FN NON-DRAGEN-FP False negatives and false positives 40,000 87% Error Reduction 33,090 DRAGEN v4.4 cont
Slide 12: DRAGEN Germline Pipeline
DRAGEN Germline Pipeline For Research Use Only. Not for use in diagnostic procedures.
Slide 13: Multigenome mapper resolves difficult-to-map regions
DRAGEN multigenome mapper resolves difficult-to-map regions and increases number of variants discovered Available since DRAGEN v3.7.5, updated in v4.4.4 with 128 Global pangenome reference * and increased SV support Multigenome Mapping with Population Haplotypes v3.4 - v3.6 v3.7 – v3.8 v3.9 v3.10 v4
Slide 14: Major leap in structural variant calling accuracy
Major leap in structural variant calling accuracy and improvements in personalized pangenome accuracy Redefining SV calling accuracy Personalized pangenome accuracy DRAGEN 4.4 improves on F-score by more than 30%* Personalized pangenome references now generally available (GA) Structural variant popu
Slide 15: DRAGEN SV with pangenome reference
DRAGEN SV with pangenome reference SV Accuracy • DRAGEN 4.4 improves on F-score by more Truth Set Method Recall Precision Fscore than 30% on HG002 with both NIST T2TQ100 1 and CMRG2 benchmarks. NIST DRAGEN 4.4 0.807 0.963 0.878 T2TQ100 • Significant recall gains in SV insertion detection HG002 DRAGE
Slide 16: Accuracy improvements of DRAGEN SV calling
Accuracy improvements of DRAGEN SV calling HG002 NIST v1.01 1 - DRAGEN SV Accuracy Second-gen multigenome mapper with new pangenome reference 1.0 0.95 0.96 0.96 0.95 0.95 0.95 0.96 0.9 0.88 0.81 0.8 0.7 0.65 0.67 0.61 0.62 0.6 0.58 0.49 0.51 0.5 0.44 0.46 0.41 0.42 0.4 0.3 0.27 0.2 0.1 0.0 1 – Compa
Slide 17: Personalized germline small variant caller
Personalized germline small variant caller Official release Leverages 128-sample pangenome reference and Activated with multigenome-mapper output to compute --enable-personalization=true personalized 2-haplotypes reference Run time on 35x WGS approx. 22 minutes for alignment ✓ Use the 2-hap personal
Slide 18: Personalized germline small variant caller – accuracy
Personalized germline small variant caller – accuracy DRAGEN v4.4 adds further small VC gain (5.35% in SNP, 10.30% in Indel). Personalization adds a further gain of 21.63% for very little added run time (3 minutes and 40 seconds). FPs+FNs, SNPs+Indels, HG002-HG007*, NIST v4.2.1 high confidence regio
Slide 19: Personalized small variant caller improvement in the dark regions
Personalized small variant caller improvement in the dark regions of the genome Personalization adds a further gain of 24.18% for SNPs and INDELS on the dark regions of the genome FPs+FNs, SNPs and INDELs, HG002-HG007*, NIST v4.2.1 difficult-to-map regions Difficult-to-map regions1 7,000 5,816 5,794
Slide 20: Personalized small variant caller for CMRG
Personalized small variant caller for CMRG Personalization reduces by 18% FPs+FNs in Challenging Medically Relevant Genes (CMRG) FPs+FNs, SNPs and INDELs, HG002*, NIST CMRG* 500 CMRG1 419 False negatives and false positives 405 400 Curated benchmark of 273 genes 137 332 132 with repetitiveness and 3
Slide 21: In-run construction of CNV panel-of-normals for germline enrichment
In-run construction of CNV panel-of-normals for germline enrichment Available via single-click GUI element, or CLI DRAGEN v4.4 provides workflow-level support for automatically constructing a PoN from a batch of samples. • Ensures robustness to batch-level effects • Reduces putative false positives
Slide 22: Germline WGS cytogenetics
Germline WGS cytogenetics B-allele frequency support for Germline CNV Support for WGS-based cytogenetics analysis • Enables the detection of AOH/LOH regions across the genome Example Command Line: • Increases calling robustness by considering both read coverage and dragen \ -r <HASHTABLE> \ minor al
Slide 23: WGS concordance with Cytogenetics
WGS concordance with Cytogenetics > 97% Concordance against karyotyping and CMA Supported events 1 0.973 1 0.967 0.979 0.977 DUPs ≥ 50kb, including whole- 0.938 chromosome 0.9 DELs ≥ 25kb, including whole- chromosome 0.8 AOH ≥ 500kb, including whole- chromosome 0.7 Mosaicism ≥ 20% 0.6 Coverage: 30-6
Slide 24: New WES Targeted Calling
New WES Targeted Calling • Support for HBA and SMN WES run • Requires Illumina Custom Enrichment Panel v2 specifically designed for DRAGEN Targeted Calling • Normalization using automatic copy-neutral sample identification clinical additional sample samples from a multiplexed WES run • Systematic no
Slide 25: SMN and HBA targeted callers coming to Exome 2.5 Enrichment
DRAGEN SMN and HBA targeted callers coming to Exome 2.5 Enrichment in v4.4 DRAGEN WES Targeted Caller Key Features Spike-in probes: Additional probes (~359 kbp) to cover current and prospective targeted caller regions. Panel of Normal (PoN): Specialized Exome 2.5 + spike-in PoN approaches to normali
Slide 26: DRAGEN Oncology
DRAGEN Oncology For Research Use Only. Not for use in diagnostic procedures.
Slide 27: WGS/WES provide the most comprehensive analysis among all molecular technologies
WGS/WES provide the most comprehensive analysis among all molecular technologies when powered by accurate and scalable informatics G EN ET I C ALT E RATI ON S I N T U M OR D NA / RN A B IO M A RK E RS SNVs & Copy Number Structural Loss of Fusions & Tumor Microsatellite HRD* Indels Variants Variants
Slide 28: Simplifying oncology biomarker detection with pre-configured pipelines
Simplifying oncology biomarker detection with pre-configured pipelines on premises D RA G EN TOO LK IT ( key fe atur es) S TR EA ML INE D P IPE LIN E S ( EA R LY A CC E S S ) Tumor/ Normal WGS Heme WGS Map/ Align QC CNV cfDNA WGS* MRD WGS* SNV Solid Tumor Panel* Methylation WGS* SV RNA Solid Tumor R
Slide 29: Illumina products provide all components for Heme WGS
Illumina products provide all components for Heme WGS Connected Insights* Library Prep Sequencing Secondary Analysis Insights Highly open platform compatible Powerful, high throughput Accurate, comprehensive and Enables streamlined variant library prep kits from IlluminaTM sequencing with ultra-effi
Slide 30: DRAGEN Heme WGS Analysis
DRAGEN Heme WGS Analysis SNPs & INDELs Structural Variants DUX4 Rearrangements Best in class variant calling Large structural variants including Detection of rearrangements 1 Replacement for detection for SNPs and INDELs recurrent gene fusions such as such as DUX4::IGH in difficult to conventional m
Slide 31: Out of the box Heme WGS application
Out of the box Heme WGS Hotspot region File VC Systematic Noise File Regions of Interest BED MDS B-ALL T-ALL AML application Variant of yes (early access) DRAGEN SNV interest & clinically actionable? SNVs/INDELs Nirvana Germline Deduplicate Tagging FLT3-ITDs First and only commercially FLT3 ITDs e.g
Slide 32: DRAGEN WGS Tumor-Normal Analysis
DRAGEN WGS Tumor-Normal Analysis SNPs & INDELs HLA HRD Including the ability to call Four-digit Human Leukocyte phased complex/ Antigen genotyping. 1 Replacement for variants. conventional methods CNA Biomarkers Simplified Interpretation CNV and CNAs, as well as arm level & Biomarkers including TMB
Slide 33: Out of the box WGS DNA Tumor-Normal Application
T/N Somatic Components Out of the box WGS Optional CNA PON (useful for FFPE) Germline CNV from previous step DNA Tumor-Normal Tumor CNA / Arm level / co-deletions CNA Application FASTQ Linear HT Systematic noise SNV + MNVs Hotspots DBs SNV/INDELs (early access) Annotate TMB MAP/A TMB BAM (Optional S
Slide 34: DRAGEN with Connected Insights enables streamlined analysis & interpretation
DRAGEN with Connected Insights enables streamlined analysis & interpretation for oncology Options for 'no-touch' automated workflow from sequencer to draft, research report Library Prep Input Sample Genomic DNA (Manual or Alignment and Variant Annotation and Filtering - Interpretation and Type Extra
Slide 35: DRAGEN MRD pipeline for tumor-informed WGS MRD
DRAGEN MRD pipeline for tumor-informed WGS MRD Sample Nucleic acid Whole genome Whole genome Data processing Reporting specimen extraction library prep sequencing 1 Baseline test Tumor 100x T/50x N fingerprint Tumor DNA DRAGEN T/N somatic WGS Library Prep NovaSeq variant calling & MRD Buffy coat DNA
Slide 36: Faster tumor-only SNV and TMB run time by ~25%
Faster tumor-only SNV and TMB run time by ~25% Significant speed up in VC Germline Tagging and TMB • No longer required to run full variant annotation • Tumor-only run time improvements up to 25% User Interface changes • Pass “--enable-variant-annotation true” and it will do full annotation on all D
Slide 37: Support for Tumor + Normal BAM/CRAM input to mapper
Support for Tumor + Normal BAM/CRAM input to mapper Enables tumor/normal end-to-end Tumor/Normal end-to-end Workflow analyses from BAM/CRAM input Systematic Noise File End-to-end analysis for tumor/normal is now Tumor BAM supported for all input types Map/Align Sort/Dedup SNV, Indel No longer need t
Slide 38: Support for Tumor/Normal workflow with UMI
Support for Tumor/Normal workflow with UMI DRAGEN now supports Tumor/Normal v4.3 and earlier, 3 runs input with UMI in a single run Tumor Map/align Sample + UMI Somatic workflow in Tumor/Normal mode may T-BAM Variant now include UMI collapsing for Tumor, or for Calling Normal Map/align VCF both Tumo
Slide 39: DRAGEN Multiomics
DRAGEN Multiomics For Research Use Only. Not for use in diagnostic procedures.
Slide 40: Eliminating limitations in single-cell analysis
Eliminating limitations in single-cell analysis Current Industry Constraints Expensive instrumentation and consumables Flexible Scalable Different Cells, 100 to 1M Cells Nuclei, Proteins Limited scale Illumina Rigid workflows Single Cell Prep Significant expertise required Simple Affordable Easy Wor
Slide 41: Illumina Single Cell 3' RNA Analysis available on DRAGEN
Illumina Single Cell 3’ RNA Analysis available on DRAGEN DRAGEN now supports scRNA analysis from Illumina Single Cell Prep Structure of Illumina Prep FASTQ files Output: Raw matrix, filtered matrix of cells, and full set of metrics output for each sample New molecular counting strategy utilizing bin
Slide 42: Analyze multiple scRNA preps simultaneously
Analyze multiple scRNA preps simultaneously with state-of-the-art methods and visualizations DRAGEN Single-Cell RNA Pipeline Connected Multiomics / Partek Flow Hardware acceleration provides DRAGEN reports highly accurate and visual reports DRAGEN outputs are available in Connected efficient seconda
Slide 43: DRAGEN Bulk RNA
DRAGEN Bulk RNA For Research Use Only. Not for use in diagnostic procedures.
Slide 44: Improved accuracy of RNA gene fusion detection
Improved accuracy of RNA gene fusion detection Improvement across versions and vs 3rd Parties With ultra-rapid analysis time Avg Runtime/Sample Accuracy Metrics (TP and FP) 0:12:58 0:12:19 4000 3642 0:11:31 3500 3000 0:10:05 2506 2500 0:08:38 2000 0:07:12 1500 0:05:46 0:05:22 1000 805 0:04:19 625 37
Slide 45: Assembled fusion sequences and precise breakpoints
Assembled fusion sequences and precise breakpoints for in- frame/out-of-frame determination Gene fusion sequence is assembled and Output of <prefix>.fusion_candidates.final aligned to determine precise breakpoints … LeftBreakpoint … chr2:134120290:+ RightBreakpoint … FusionSequence chr22:22922721:+
Slide 46: RNA splice variant caller – increased sensitivity
RNA splice variant caller – increased sensitivity with new ML model and mapper updates Enhanced RNA splice variant caller ✓ Updated mapper tuned for better splice sensitivity ✓ New ML model with scoring and filters ✓ Trained on WTS NIST HG005 with long read data ✓ Outputs: - Intragenic splice varian
Slide 47: DRAGEN on AWS F2
DRAGEN on AWS F2 For Research Use Only. Not for use in diagnostic procedures.
Slide 48: New AWS EC2 F2 FPGA instance types deliver faster TAT
New AWS EC2 F2 FPGA instance types deliver faster TAT and improved scalability Second-generation FPGA- DRAGEN v4.4 with native powered instance types support for F2 All BSSH/ICA apps with v4.4 New CPUs with more vCPU cores and prioritizes execution on f2 faster clock speed instances Improved turn-ar
Slide 49: DRAGEN on AWS EC2 F2 performance
DRAGEN on AWS EC2 F2 performance Significant reduction in turn-around time across all workflow types DRAGEN run times f1 vs f2 Reference hg38 alt-masked graph 260 241 240 HG002 Novaseq PCR free 35x 220 -47% SNV: FQ > MA, SNV VC, 200 WGS Targeted > BAM, gVCF out 180 160 All Callers: FQ > MA, SNV VC,
Slide 50: Toolkit and platform updates
Toolkit and platform updates
Slide 51: BCL Convert updates for greater flexibility and efficiency
BCL Convert updates for greater flexibility and efficiency Gain flexibility with Optimize memory new sample sheet settings with updated outputs OverrideReads: Redefine read structure Memory optimization: New settings by overriding RunInfo (read cycle reduce memory usage for large sample boundaries,
Slide 52: gVCF Genotyper in ICA: aggregating variants at scale
gVCF Genotyper in ICA: aggregating variants at scale From gVCF to msVCF, variant annotation and pgen output using the PopGen CLI CLI step1 shard genome CLI step2 CLI step3 recal. cohort shard subshard subshard subshard batch 1 recal. gVCF recal. gVCF gVCF gVCF ML census global census Filter pgen ful
Slide 53: ORA compression - BCL to FASTQ.ORA up to 30% faster
DRAGEN ORA compression - BCL to FASTQ.ORA up to 30% faster Example DRAGEN CMD line • Convert BCL into ORA skipping intermediate FASTQ.GZ files, and decreasing the runtime up to 30% dragen --bcl-compression-only-true --bcl-input-directory <DIR> • When using compression during BCL convert, simply add
Slide 54: Operating System support
Operating System support Out with the old, in with the new CentOS 7 support is deprecated On-premise builds, AWS AMIs and Azure VMs for el7 still exist for v4.4 but no longer distributed. Customers in need may ask for the el7 builds. Guidance that v4.5 will not have the ability to generate el7 anymo
Slide 55: DRAGEN unified JSON metrics
DRAGEN unified JSON metrics Unified JSON output facilitates use of metrics of interest in downstream analysis Sample output JSON output: Unified JSON Metrics "metadata": { • DRAGEN 4.4 generates a unified JSON metrics file "dragenVersion": ”4.4.123", <output_prefix>.metrics.json for every run "licen
Slide 56: BYOL Cloud – Now you can view your licenses
BYOL Cloud – Now you can view your licenses • Can retrieve license information utilizing the packaged dragen_lic tool for BYOL Cloud credentials. • Provide the dragen_lic tool the credentials using the same DRAGEN commands, --lic-server or --lic-credentials. • Can export to a JSON format using the –
Slide 57: New DRAGEN v4.4 BSSH Apps & ICAv2 Pipelines
New DRAGEN v4.4 BSSH Apps & ICAv2 Pipelines BSSH Apps (ICAv2 backend) ICAv2 Pipelines (w/ F2 support) • BCL Convert • BCL Convert • Germline • Germline (WGS & Enrichment) • Somatic • Somatic (WGS & Enrichment) • Enrichment • Reference Builder • Germline Enrichment from BCLs • Baseline Builder • DRAG
Slide 58: Request a DRAGEN free trial
Request a DRAGEN free trial Accurate, comprehensive, efficient secondary analysis For Research Use Only. Not for use in diagnostic procedures. 58
Slide 59: Thank you!
Thank you! Illumina’
Slide 60: Appendix
Appendix Illumina’
Slide 61: New DRAGEN app manager CLI
New DRAGEN app manager CLI Streamlined DRAGEN interface to install, manage and execute comprehensive DRAGEN apps Flexible deployment options DRAGEN App Manager CLI with standalone app packages Inputs Reference FASTQs Genome Bed, VCF, Easily install Nextflow DRAGEN workflows as etc... single app pack
Slide 62: DRAGEN CNV updates
DRAGEN CNV updates Somatic CNV Model Improvements • Updates to purity/ploidy estimation allowing for higher success rate of model detection on challenging tumors. • Improved detection for samples with low coverage, very few CNVs, noisy BAF profiles, FFPE, or extreme GC bias. WGS Panel of Normals Sup
Slide 63: Improved sensitivity for low-VAF variants in amplified regions
Improved sensitivity for low-VAF variants in amplified regions with sample type auto - detection • Column-wise detection update to identify candidate variant events in high-coverage regions • Improved identification of very low VAF events in copy number amplifications • With new auto sample type det
Slide 64: Germline small variant caller – Accuracy v4.2.1 SNPs
Germline small variant caller – Accuracy v4.2.1 SNPs DRAGEN 4.4 variant caller improvements result in 5.35% reduction in SNPs FP+FN Average SNPs FP+FN reduction: 5.35% FPs+FNs, SNPs, NIST v4.2.1 high confidence regions* 7,000 5,866 5,880 5,900 6,000 5,692 5,660 5,748 5,515 5,347 5,220 False negative
Slide 65: Germline small variant caller in the dark regions – SNPs
Germline small variant caller in the dark regions of the genome 5.58% reduction in SNPs FP+FN in difficult-to-map regions FPs+FNs, SNPs, HG002-HG007*, NIST v4.2.1 difficult-to-map regions FP FN Difficult-to-map regions1 6,000 regions that have other 5,352 5,354 5,391 5,150 5,224 5,158 homologous reg
Slide 66: Germline small variant caller in the dark regions – INDELs
Germline small variant caller in the dark regions of the genome 4.39% reduction in INDELs FP+FN in difficult-to-map regions FPs+FNs, INDELs, HG002-HG007*, NIST v4.2.1 difficult-to-map regions FP FN Difficult-to-map regions1 500 regions that have other 458 440 450 425 homologous regions in the 411 41
Slide 67: New cytogenetics output reduces large call fragmentation
New cytogenetics output reduces large call fragmentation DUP / DEL / AOH CNV+SV VCF CYTO VCF (*.cnv_sv.vcf.gz) (*.cyto.vcf.gz) [1kb, 25kb) Supported N/A [25kb, 500kb) Supported Recommended (RES=25k) [500kb, 1Mb) Supported Recommended (RES=500k) ≥ 1Mb Supported Recommended (RES=1M) Mosaic Alterations
Slide 68: Germline CNV can now detect AOH events
Germline CNV can now detect AOH events > 97% concordance with CMA results on internal AOH validation set (N=34 samples) Degree of autosomal homozygosity reported by DRAGEN used as Consanguinity estimate Sample Metric Expected Estimated HG01970 Autosomal AOH % (hg38) 12.2 11 CMA truth NA19679 Autosom
Slide 69: Cytogenetics modality provides support for multi-resolution views
Cytogenetics modality provides support for multi-resolution views Poo r Poo r mappa bility mappa bility reg ions reg ions Target Target counts counts coverage coverage B-Allele B-Allele freque ncy freque ncy BAM BAM coverage coverage BAM read s BAM read s distribution distribution CMA truth CMA trut
Slide 70: DRAGEN v4.4 supports genotyping 41 HLA and HLA-related genes
DRAGEN v4.4 supports genotyping 41 HLA and HLA-related genes Nucleotide-based alignments using DRAGEN map-align and expectation-maximization for HLA genotyping Extract HLA reads: BED filter Process and filter alignments based on DRAGEN alignment against (auto-detected reference: DRAGEN alignment sco
Slide 71: HLA genotyping accuracy
HLA genotyping accuracy Class I genes (A, B, and C), Class II (DQB1, DRB1) 1KG truth set Class I genes (A, B, and C), Optitype truth set DRAGEN v4.4 HLA caller shows highest accuracy for Class I DRAGEN v4.4 HLA caller shows higher accuracy overall and higher accuracy genes (A, B, and C) compared to
Slide 72: DRAGEN v4.4 HLA supported genes and concordance against T1K
DRAGEN v4.4 HLA supported genes and concordance against T1K DRAGEN concordance against T1k over 3,202 diverse WGS samples from 1KGP gene gene type matches N %Concordance gene gene type matches N %Concordance A Class I 6373 6404 99.52 F Class I minor 6368 6404 99.44 B Class I 6384 6404 99.69 G Class
Slide 73: WES targeted calling from command line
WES targeted calling from command line • Three steps for targeted calling from a WES run • Available as a single workflow via ICA/BSSH app • For correct normalization, each WES run must be processed as a batch and include samples from a single library prep. • Counts generation: For each sample in a
Slide 74: STR motif composition for C1 calling
STR motif composition for C1 calling (and beyond …) Catalog specification example • Pathogenicity of STR loci like RFC1 is highly (HMM decomposition) { dependent on the motif composition "LocusId": "RFC1", "LocusStructure": "(AARRG)*", • DRAGEN 4.4 introduces STR motif composition: "ReferenceRegion"
Slide 75: STR motif composition estimation helps identify pathogenic motif in RFC1
STR motif composition estimation helps identify pathogenic motif in RFC1 locus in individuals affected by CANVAS STR sizes and pathogenic motif fractions estimated by DRAGEN STR v4.4 Number of repeat units haplotype 1 Number of repeat units haplotype 2 Pathogenic motif (AAGGG) fraction CANVAS positi
Slide 76: DRAGEN Single-Cell RNA workflow
DRAGEN Single-Cell RNA workflow Per-read processing RNA Alignment FASTQs Sorted DBAMs with Correct barcode/BI blobs barcodes per RNA read Cell by Map/align, feature sorting matrix GTF Barcode counting and correction Match reads to Cell filtering Barcode / BI genes and parsing and features dynamic ph
Slide 77: Streamlined multi-region joint detection (MRJD)
DRAGEN v4.4 - streamlined multi-region joint detection (MRJD) for de novo germline small variant calling in paralogous regions MRJD and DRAGEN Small Variant Caller can now be enabled in the same DRAGEN run Covers six clinically relevant genes - NEB, TTN, SMN1/2, PMS2, STRC, and IKBKG New tags in the
Slide 78: DRAGEN MRD: Demonstrated competitive analytical performance
DRAGEN MRD: Demonstrated competitive analytical performance Cancer Type Novaseq 6000 Novaseq X NSCLC 0.003% 0.003% CRC 0.005% 0.003% LoD95 Bladder 0.005% 0.005% Melanoma 0.006% 0.006% Breast 0.006% 0.006% 100% 100% Analytical Specificity [95% CI: 99.1%-100%] [95% CI: 99.1%-100%] • Analytical sensi
Slide 79: Updated gene fusion + splice variant caller interfaces
Updated gene fusion + splice variant caller interfaces • Normally only passing intergenic splice variants are passed to the fusion caller to be scored by the fusion ML model (skips filtered or readthrough genes) • Option added to specify a list of hotspot genes that may contain spliced fusions: --rn
Slide 80: Example of read-through fusion calling (SLC45A3-ELK4)
Example of read-through fusion calling (SLC45A3-ELK4) Entry in intergenic splice variants.csv in v4.4 • SLC45A3-ELK4 RNA read-through fusion (cis splicing of neighboring gene_start ELK4 genes) is an onco-marker for prostate cancer(+) gene_end SLC45A3 chromosome chr1 • Called (PASS) in v4.4 with --rn
Slide 81: MNV calling enabled by default in somatic small VC
MNV calling enabled by default in somatic small VC • By default, phased variants within 2bp of one another will be Sample MNV record and component call records filtered as mnv_component: merged and reported as an MNV/delins • Follows HGVS guidelines for variant reporting and enables better downstrea
Slide 82: Illumina Connected Annotations - Data Manager
Illumina Connected Annotations - Data Manager • New utility to give the customer complete control over annotation data. • Available functionalities: • View available versions. • Create data configuration template. • Download data. • Validate data versions. https://illumina.github.io/IlluminaConnecte
Slide 83: Illumina Connected Annotations - ISCN like nomenclature
Illumina Connected Annotations - ISCN like nomenclature • Provides ISCN-like simple nomenclature to describe karyotype of the input in sample's and variant's level in both VCF and JSON outputs. Chromosome Start Position End Position Variant Type ISCN Notation 8 19200001 135400001 deletion del(8)(p21
Slide 84: Flexible gVCF metric import for iGG
Flexible gVCF metric import for iGG To set the fields to be ingested: Pick the metrics you need from the input gVCFs --gg-format-to-import=… --gg-info-to-import=… INFO fields available for import: Allows import of metrics beyond the default set MQ, MQRankSum, ReadPosRankSum, (which is unchanged). Al
Slide 85: Streamlined workflow for multi-batch msVCF file output
Streamlined workflow for multi-batch msVCF file output Go directly from cohort files to all-sample msVCF Step 1 gvcf gvcf gvcf gvcf gvcf gvcf … gvcf gvcf gvcf aggregate gVCFs cohort1 cohort2 cohortN census1 census2 censusN Step 2 aggregate census global census (site) (call site) New multi-cohort Ste
Slide 86: Germline small variant caller – Accuracy v4.2.1 indels
Germline small variant valler – Accuracy v4.2.1 indels DRAGEN 4.4 variant caller improvements results in 10.30% reduction in Indels FP+FN Average Indel FP+FN reduction: 10.30% FPs+FNs, Indels, NIST v4.2.1 high confidence regions* FP FN 3,000 2,425 2,500 False negatives and false positives 2,284 2,25
Slide 87: DRAGEN ORA compression
DRAGEN ORA compression Store and easily retrieve information about your sample • From BCLconvert to FASTQ.ORA, additional metadata are Example DRAGEN CMD line to generate the automatically stored to FASTQ.ORA files metadata json file • The command --ora-get-metadata. true, generates a Json file with
Slide 88: ORA compression - BCL to FASTQ.ORA up to 30% faster
DRAGEN ORA compression - BCL to FASTQ.ORA up to 30% faster Runtime BCLconvert to FASTQ.ORA with DRAGEN 4.4 dataset: 25B flowcell data Runtime BCL to ORA (min) per lane 78 Runtime (min) 53 Without option With option --bcl-ora-direct --bcl-ora-direct Test conducted on DRAGEN server v4 with all other s
Slide 89: CRAM 3.1 support
CRAM 3.1 support 30x WGS HG002 hg38 DRAGEN now supports CRAM 3.1 16 14.22 14 12.83 Input or generate CRAM files with format 3.1 12 File Size (GB) 10 8 CRAM 3.1 files are 7-15% smaller than CRAM 3.0 6 4 2 No impact on run time 0 CRAM 3.0 CRAM 3.1 New option --cram-version=3.1 (default is 3.0) Availab
Slide 90: New license server domain
New license server domain – license.dragen.illumina.com On-premises servers BYOL Cloud • HTTPS @ license.edicogenome.com will be • HTTP @ lus.edicogenome.com is being replaced with HTTPS @ replaced with HTTPS @ license.dragen.illumina.com license.dragen.illumina.com • All versions of DRAGEN support
Slide 91: Fractional downsampler feature for RNA Pipelines
Fractional downsampler feature for RNA Pipelines Simulates different amount of sequencing for high coverage samples To enable the fractional downsampler: • Fractional downsampling can be utilized with RNA pipelines --enable-fractional-down-sampler=true • Subsampling based on user-defined percentage
Slide 92: DRAGEN Checkfingerprint
DRAGEN Checkfingerprint Determine if the input sequencing data originates from the same sample Input observed sample (bam/vcf/fastq) Input expected sample DRAGEN DRAGEN Checkfingerprint Checkfingerprint Output (vcf) Users are required to provide inputs for both the observed Pre-computed sample and t
Slide 93: DRAGEN Checkfingerprint
DRAGEN Checkfingerprint To enable the Checkfingerprint: • Checkfingerprint compute identity between input samples against a set of pre-computed genotypes. --enable-checkfingerprint true --checkfingerprint-expected-vcf • The primary output is LOD score, which indicates the [path_to_expected_sample_vc
Slide 94: DRAGEN reports
DRAGEN reports Key benefit of updates 16S Report • Replacement for one of the oldest and most widely used legacy BaseSpace apps • Added multi-column reports for more complicated designs and smaller plots and tables • Adds support for multiple new plot types, such as Sunburst plots and Dendrograms Ad
Slide 95: DRAGEN BSSH and ICA Apps Updates
DRAGEN BSSH and ICA Apps Updates For Research Use Only. Not for use in diagnostic procedures.
Slide 96: BCL Convert
BCL Convert Key benefit of updates Non-DRAGEN features or bug fixes • Allow dots and space characters in RunName and ProjectName (consistent with Run Planning) 96 For Research Use Only. Not for use in diagnostic procedures.
Slide 97: DRAGEN Germline
DRAGEN Germline Key benefit of updates New DRAGEN features User interface changes • Updated to DRAGEN 4.4.4 • Added support for BAIs and CRAIs • Support v11 reference hash tables (4.4 feature) • Added additional hs37d5 reference with chr prefix • Added support for CRAM v3.0 and CRAM v3.1 input • Cle
Slide 98: DRAGEN Enrichment
DRAGEN Enrichment Key benefit of updates New features User interface changes • Upgraded to DRAGEN 4.4.4 • Support v11 reference hash tables (4.4 feature) • Added additional hs37d5 reference with chr prefix • Removed under-utilized options • Added support for CRAM v3.0 and CRAM v3.1 output • Added su
Slide 99: DRAGEN Somatic
DRAGEN Somatic Key benefit of updates New features User interface changes • Upgraded to DRAGEN 4.4.4 • Support v11 reference hash tables (4.4 feature) • Added additional hs37d5 reference with chr prefix • Added support for CRAM v3.0 and CRAM v3.1 output • Enabled UMI support for Tumor-Normal analysi
Slide 100: DRAGEN RNA
DRAGEN RNA Key benefit of updates New DRAGEN features User interface changes • Upgraded to 4.4.4 • Support splice variant calling • Support fractional downsampler • Support v11 reference hash tables (4.4 feature) • Support built-in mm39 and hg38-mm39 references • Set annotation genome automatically
Slide 101: DRAGEN Amplicon
DRAGEN Amplicon Key benefit of updates New DRAGEN features User interface changes • CheckFingerprint always run in VCF comparison mode. • Support v11 reference hash tables (4.4 feature). • Support CRAM v3.1. Non-DRAGEN features • ICAv2 backend for the BaseSpace app • Auto-detect annotation genome. •
Slide 102: DRAGEN Microbial Amplicon (DMA)
DRAGEN Microbial Amplicon (DMA) Generates consensus sequences from Illumina Microbial Amplicon Prep (iMAP), COVIDSeq, or other amplicon-based libraries from microbial samples Key benefit of updates v1.1.0 release notes DMA replaces BSSH apps near end of life • Updated DRAGEN to 4.4 • DRAGEN Targeted
Slide 103: DRAGEN joint pedigree and E2E joint pedigree
DRAGEN joint pedigree and E2E joint pedigree Key benefit of updates Overview • 2 apps, one starting from GVCFs and another starting from FASTQs • Processes one family (pedigree) at a time Features • Update to DRAGEN 4.4.4 • Add REViewer—a viewer for repeat expansions 103 For Research Use Only. Not f
Slide 104: DRAGEN Reference Builder
DRAGEN Reference Builder Key benefit of updates New DRAGEN features • Generate references in the v11 hash table format 104 For Research Use Only. Not for use in diagnostic procedures.
Slide 105: DRAGEN Population Haplotyping
DRAGEN Population Haplotyping ICA Key benefit of updates New features User interface changes • Upgraded to DRAGEN 4.4.4 • Added PassOnly filter option during phase common step Non-DRAGEN features or bug fixes • Supports contig-specific joint-VCF input • Reduced input requirements (no reference FASTA
Slide 106: DRAGEN Imputation Reference Panel Builder
DRAGEN Imputation Reference Panel Builder ICA Key benefit of updates New DRAGEN features User interface • Packs the per-contig output from the DRAGEN Population Haplotyping pipeline into a custom imputation reference panel TAR.GZ that can be used with DRAGEN Imputation pipeline or download. • Suppor
Slide 107: DRAGEN Baseline Builder
DRAGEN Baseline Builder Key benefit of updates New DRAGEN features User interface changes • Upgrade to 4.4.4 • Support v11 reference hash tables (4.4 feature) • Split support for CNV WGS and CNV Enrichment • Always creates combined counts file in CNV modes Non-DRAGEN features • Support setting ICA s