Releases: hartwigmedical/pipeline5
v5.33
Pipeline v5.33 Release Notes
Summary of improvements:
-
RefSeq gene annotations:
Variants in certain genes are now annotated with RefSeq transcript IDs in addition to the Ensembl transcript IDs. -
Improved targeted panel analysis:
Various tools contain updates to support targeted panel analysis better.
Detailed Tool Changes
| Component | Version | Key Functional Changes |
|---|---|---|
| Purple | 3.9 | Various panel changes, add somatic likelihood |
| Sage | 3.3 | Track UMI counts, Gene coverage to min map quality >= 10, minor panel filtering changes |
| Pave | 1.5 | Phased variant fixes |
| Cobalt | 1.15.2 | Panel GC refinements & fixes |
| BamTools | 1.1 | Exclude duplicates from fragment length distribution |
| MarkDups | 1.0 | Initial version with read consensus & UMIs |
| Linx | 1.24.1 | Visualiser improvements and fusion plot fixes |
| Lilac | 1.5 | Output as TSV |
| PatientDB | 5.33 | Add Clinvar, Gnomad and Somatic Likelihood to somatic variant table |
| Cuppa | 1.8.1 | Alternative SJ fix for RNA samples |
| Health-Checker | 3.5 | Compatibility with samtools Flagstats (instead of from Sambamba) |
| Virus Interpreter | 1.3 | Use updated virus enumerations |
| Orange | 2.6.0 | Support non standard transcripts from PURPLE vcf |
| Chord | 2.02 | Require 100 INDELs or 30 SVs to make a prediction, as opposed to 50 INDELs or 30 SVs before |
Changes to resource files
| File | Tool | Change |
|---|---|---|
| Driver gene panel | Many | Several germline genes added |
| CoverageCodingPanel | Sage | All driver panel genes included |
| Ensembl data cache | Pave | RefSeq transcripts added |
| Fusions | Linx | FGFR1 HighImpactPromiscuous back to false |
| Clinvar | Pave | Updated to latest |
| KnownGermlineHotspots | Sage | Following Clinvar update |
| KnownSomaticHotspots | Sage | Six manually curated hotspots added |
| HMF targeted panel files | Purple, Cobalt | Normalisation to use 64 samples, TMB calc adjustments |
| Old Gridss resource files | Gridss | Deleted/cleaned up |
v5.32
Pipeline v5.32 Release Notes
v5.32 has these key improvements:
-
Germline SV reporting – annotate germline SVs with tumor copy
number info -
Purple purity enhancements – somatic fit improvements
-
Promiscuous enhancer fusion calling for TERT and C19MC
For more information see:
Detailed Tool Changes
| Component | Key Functional Changes | Full Release Notes |
| Purple (3.7-> 3.8.2) | - lower somatic fit allele count threshold from 2K -> 1K - write germline SV VCF with copy number annotations for Linx Germline - set germline deletion positions to match germline SVs where possible - calculate hotspot VAF when somatic peak isn't established - add probability test for biallelic if MACN > 0.5 - plot somatic variant regardless of SNV count but only plot first 100K - increase somatic weight penalty 1 -> 1.5 - somatic fitting peak variants 10 -> 4 - exclude hotspots from somatic fitting filters - cap germline CN at 1 for germline status HET_DELETION in somatic variant CN - revert from somatic to normal fit if deleted depth window percent exceeds 0.3% - trisomy cutoff 1.4 -> 1.35 - Amber read-count range 0.6-1.4 -> 0.8-1.2, and exclude hotspots from read-count range - extra criteria on fitting variants: germline diploid, zero ref allele count, mappability 1, repeat count < 4, not in Gnomad, not low confidence - driver gene panel reportGermlineDisruption now a Wild/Variant/Any/None type field - write all germline deletions not just for driver genes - ensure VCF sample IDs match config in name and order |
|
| SvPrep(1.0.1 -> 1.1) | - exit on any error within a thread - improved depth annotation memory and performance |
1.1 |
| Gripss (2.3.2 -> 2.3.5) | - provide tumor sample for germline mode and 'germline' flag, so as to write tumor genotype info the VCF (and for Linx to use) - use tumor ID in output VCF file name in germline mode |
2.3.5 |
| Sage (3.2.3 -> 3.2.5) | - adapter filter applied earlier, and proximate DEL for soft-clip extension tightened - add missing ABQ VCF header tag if missing for Sage append mode - improper pair penalty only applied to paired reads - ignore hotspots in min avg base qual soft filter - fixed max proximate DEL as used in soft-clip core overlap search |
3.2.4 3.2.5 |
| Linx (1.22 -> 1.23.2) | - new fusion type PROMISCUOUS_ENHANCER_TARGET, non-reportable with any breakend in 3' gene/region, must be > 100K and not a SGL/INF - disruptions set biallelic, max/min copy number and likelihood method - IG known pair downstream genes consider breakends up to 250K upstream - populate tumor fragments in germline mode - supress linked SGL-INF breakends from consideration in gene routines - fragile site and line source element files added as internal resources, also used in the Visualiser - process and write non-canonical drivers to driver catalog file - check closed loop DM chain's last link in fusion finder - added SvId and VcfId to sv germline file and database - germline: add driver type 'GERMLINE_HOM_DUP_DISRUPTION' for germline homozygous DUP disruptions - germline: use Purple CN info to set germline undisrupted values - germline: write breakend data, removed gene from disruption file and table - set SGL-INF type using mapping position for DELs and DUPs - visualiser: skip fusion component plot if either up or down gene exons are missing - visualiser: fixed protein domain bug - visualiser: switched to use CytoBands from common, add hg38 file for visualiser - visualiser: fixed crash when fusion has no overlapping upstream exons |
|
| Pave (1.4 -> 1.4.3) | - factor out superfluous ref codon bases for conservative INDEL overlap test in phasing logic | |
| Cuppa (1.4 -> 1.8) | - generate R PDF report from within Cuppa - support GRCh38 |
1.7.2, 1.8 |
| BamTools (1.1) | Initial version – replaces call to Picard CollectWgsMetrics | 1.0 |
Changes to resource files
| File | tool | Change |
| Fusion Reference | Linx | Promiscuous Enhancers TERT and C19MC PROMISCUOUS_3 TERT KNOWN_PAIR BRD4 LEUTX KNOWN_PAIR MYBL1 C8orf34 IG_KNOWN_PAIR IGL ETV6 KNOWN_PAIR PLAGL1 FOXO1 KNOWN_PAIR DEK AFF2 |
| Driver Gene Panel | Added germline disruption and deletion types |
v5.31
Key improvements:
Improved performance of our SV caller, GRIDSS. A new component ‘SV Prep’ has been introduced which both pre-filters the BAM prior to GRIDSS assembly and also implements faster depth annotation post assembly. This reduces GRIDSS runtime by ~70%. Minor changes have also been made to GRIPSS filters associated with this change
Improved support for targeted sequencing. The TMB/TML/MSI models have been completely overhauled in PURPLE for targeted sequencing as well as better filtering rules for AMP and DEL drivers Changes. In SAGE a new filter has been added and the BQR logic has been altered to better handle FFPE samples we are using in targeted analysis
Tool upgrades:
| Component | Key Functional Changes | Full Release Notes |
|---|---|---|
| PURPLE (3.5->3.7.1) | Global changes • removed support for SnpEff enrichment (PAVE required) • Check MNVs and INDELs hotspots for NO_TUMOR check (previously only SNV) • Somatic variant allele read count threshold lowered for NO_TUMOR check (5k->2k) • driverLikelihood = 0 if gene not in DNDS data (previously PURPLE failed) • CN smoothing abs CN diff threshold lowered from 2 to 1 (ie less smoothing) • CN region start position set to mid of min-max start (previously set to min) • BUG FIX: CN min start cannot overlap previous segment start position Targeted mode specific • New MSI model for targeted mode • New TML & TMB model for targeted mode • DEL drivers require 3+ depth windows in targeted mode • PARTIAL_AMP drivers only in genes with known pathogenic exon deletions • TR: allow partial amps for specific genes (PROVIDE LIST) • TR: apply AF and germline likelihood filter for TMB | 3.6, 3.7, 3.7.1 |
| SV PREP (1.0.1) | => New component | 1.0, 1.0.1 |
| GRIPSS (2.2->2.3) | • DiscordantPairSupport filter now only applies to short INV • Short INV threshold changed from 40 to 50 bases | 2.3 |
| SAGE (3.1 -> 3.2.3) | • New ‘minAverageBaseQual’ filter (primarily for FFPE BQR uses panel regions (to support targeted NGS) and allows higher rates of errors at individual bases • New exon median depth coverage output (for QC reporting) • Soft clipping ignored for variant calling where fragment length =~ mapped bases (likely adapter sequence) | 3.2 3.2.2, 3.2.3 |
| LINX (1.20 -> 1.22) | • Visualiser: allow genes with clusterIds in config when gene is linked to another cluster • Added 'germline' to annotation, cluster and link files in germline mode • No duplicate disruption records added to driver catalog | 1.21, 1.22 |
| PAVE (1.3->1.4) | • BUG FIX: HGVS coding for enhancer MNV (TERT) | 1.4 |
| LILAC (1.1->1.4) | • somatic variant calling does not depend on gene impact annotation • choose non-wildcard over wildcard allele for somatic variants | 1.2, 1.3, 1.4 |
| PEACH (1.6->1.7) | • Support UGT1A1 haplotype calling | 1.7 |
Resource File Changes
| HMF panel cobalt normalisation file | Cobalt (panel only) | Low enrichment regions masked (<0.1 median enrichment) |
|---|---|---|
| Driver Gene Panel | enable germline calling on future PGX genes: UGT1A8, CYP3A4, CYP3A5 and CYP2D6 | |
| Gnomad vcfs | (Pave) | Now includes exome and genome data merged (hg38 only) |
| sage/38/PanelArtefacts.38.tsv | (PAVe – panel only) | Updated from PON built from all HMF panel samples + manual excluding normal PON |
| peach/min_DPYD.json -> peach/peach.json | PEACH | Add UGT1A1 configuration and rename file |
v5.30
- Upgrade Protect 2.3
- Upgrade Rose 1.1
- Upgrade Linx 1.20.1
- Upgrade Orange 1.10.1
- Upgrade Purple 3.5.1
- Upgrade Cuppa 1.7.1
- Upgrade Sage 3.1.1
- Upgrade SERVE 1.12
- Persist Lilac sliced bam for faster reruns
- Migrate from pre-emptible to spot vms
- Add DPYD to sage panel files
- Add report PGX flag to gene panel
- Expand CCND1 3' UTR calling
Pipeline Docker Image: 5.30.1
Pipeline VM Image: pipeline5-5-30-202208221829-202208240943-private
v5.29
- Add ROSE (version 1.0) to pipeline
- Separate BAM slicing step from the rest of the Lilac stage
- Separate virus breakend and interpreter steps
- Update driver gene panel and fusion
- Upgrade ORANGE to 1.10
- Upgrade PAVE to 1.2.2
- Upgrade PROTECT to 2.2
- Upgrade SERVE to 1.11
- Remove potentially pathogenic germline variants from GRIDSS PON
- Switch to using API to determine lanes rather than parsing FASTQ filename
Pipeline Docker Image: 5.29.1
Pipeline VM Image: pipeline5-5-29-202206241631-202206302035-private
v5.28
- Support for running with tumor only and germline only
- Support for narrowing analysis to target regions (ie panel/enrichment)
- Upgrade SAGE 3.0
- Upgrade PAVE 1.2
- Upgrade COBALT 1.13
- Upgrade GRIPSS 2.1
- Upgrade HEALTH_CHECKER 3.4
- Upgrade LINX 1.19
- Upgrade AMBER 3.9
- Upgrade PURPLE 3.4
- Upgrade SERVE 1.10
- Upgrade PROTECT 2.1
- Upgrade ORANGE 1.7
- Upgrade PEACH 1.6
Pipeline Docker Image: 5.28.1
Pipeline VM image: pipeline5-5-28-202204281626-202204281749-private
v5.27
- Add Germline SV and CN calling with LINX, GRIPSS and PURPLE
- Disable somatic reporting for DPYD
- Add LILAC 1.1
- Upgrade GRIDSS 2.13.2
- Upgrade AMBER 3.6
- Upgrade Health Checker 3.3
- Upgrade ORANGE 1.6
- Upgrade PROTECT 2.0
- Upgrade Virus Interpreter 1.2
- Upgrade PURPLE 3.3
- Upgrade PAVE 1.1
- Upgrade LINX 1.18
Pipeline Docker Image: 5.27.1
Pipeline VM image: pipeline5-5-27-202202221543-202202221631-private
v5.26
Features:
- Add PAVE 1.0 https://github.com/hartwigmedical/hmftools/releases/tag/pave-v1.0
- Upgrade PEACH 1.5
- Upgrade GRIPSS 2.0 https://github.com/hartwigmedical/hmftools/releases/tag/gripss-v2.0
- Upgrade PURPLE 3.2 https://github.com/hartwigmedical/hmftools/releases/tag/purple-v3.2
- Upgrade CUPPA 1.6 https://github.com/hartwigmedical/hmftools/releases/tag/cuppa-v1.6
- Upgrade LINX 1.17 https://github.com/hartwigmedical/hmftools/releases/tag/linx-v1.17
- Upgrade ORANGE 1.5 https://github.com/hartwigmedical/hmftools/releases/tag/orange-v1.5
- Upgrade PROTECT 1.9 https://github.com/hartwigmedical/hmftools/releases/edit/protect-v1.9
- Upgrade SERVE 1.8 https://github.com/hartwigmedical/hmftools/releases/tag/serve-v1.8
Resources:
- Update of Ensembl, driver and fusion genes to match HGNC
- Removal of SnpEff config (no longer used)
- Removal of Linx viral host files and replication origins file
- Sorted GRIPSS PONs
- Sorted driver gene panels on gene names
- Serve 1.8 resources
v5.25
Features:
- Upgrade Virus Interpreter v1.1: https://github.com/hartwigmedical/hmftools/releases/tag/virus-interpreter-v1.1
- Upgrade CUPPA to v1.5: https://github.com/hartwigmedical/hmftools/releases/tag/cuppa-v1.5
- Upgrade PROTECT to v1.5 and 1.6: https://github.com/hartwigmedical/hmftools/releases/tag/protect-v1.5 https://github.com/hartwigmedical/hmftools/releases/tag/protect-v1.6
- Upgrade ORANGE to v1.2-v1.4 https://github.com/hartwigmedical/hmftools/releases/tag/orange-v1.2 https://github.com/hartwigmedical/hmftools/releases/tag/orange-v1.3 https://github.com/hartwigmedical/hmftools/releases/tag/orange-v1.4
- Consolidate analysis types into a single event (HMF infrastructure change)
- Add disclaimer to CUPPA report output
- Support separate cuppa feature plot for ORANGE
Resource Updates:
- Updated SERVE resources
- Remove germline pathogenic PON entries
- Driver Gene Panel updated
- Fusion knowledgebase updated
v5.24
Features
- Added ORANGE v1.1: https://github.com/hartwigmedical/hmftools/releases/tag/orange-v1.1
- Upgrade PEACH to v1.3: https://github.com/hartwigmedical/peach/releases/tag/v1.3
- Add CUPPA report PNG to pipeline output
Resource Updates:
- Add missing cup ref snv signatures to cuppa resources
- Update SERVE to 1.6 https://github.com/hartwigmedical/hmftools/releases/tag/serve-v1.6
- Update germline hotspots including a new blacklist