Releases: hartwigmedical/hmftools
Releases · hartwigmedical/hmftools
sv-linx v1.5
Functional changes:
- Fusion likelihood calcs - implemented gene-pairs, skip non-coding same-gene fusions
- known fusions don't require upstream transcript to be disruptive but will still fail on chain termination
- TIs limited by min of achor distances now less sum of breakend homologies
- BFB_AMP does not require max ploidy x8
- dominant foldback must exceed ploidy 4
- DMs can have a single SV and need to be 5x sample ploidy or 2.3x telomere/centromere
- dissolve simple groups unless all SVs in LOH or all assembled
- set intronic transcripts from LINE clusters to be non-disruptive
Bugs:
- only reconfigure chains for RECIP_INV_DUPs if can form positive TI
- determination of unique non-reported fusions was skipping first element
amber-v3.0
Added support for tumor-only mode.
Replaced bed file and snp check file with VCF
Bachelor v1.9
bachelor-v1.9 Single process and pipeline v5 enabled
sv-linx-v1.4
- Disruptive definition is changed for sv breakends. Small chained templated insertions (<5k bases) into or from intronic sections are no longer considered disruptive and will not be reported in the SV section of the report
- The number of undisrupted copies is calculated for each disruptive breakend
- Homozygous disruptions are now reported as drivers by LINX and are in the driver catalog
- Each amplification, deletion or LOH driver in the driver catalog is now linked to the SVs that caused it in the svDriver table
- Breakends in the 3’UTR region of the upstream gene are permitted to form fusions via exon skipping
purple-v2.34
Added driver catalog to file output
Purity sunrise plot now supports somatic inferred purity
Additional BAF inferring rule
Removed GERMLINE_AMPLIFICATION copy number method.
Added NON_DIPLOID copy number method. This uses the ref normalised tumor copy number (adjusted for observed normal ratio rather than ideal).
Only extend long arm copy numbers to unknown regions of chromosomes 13-15,21,22
Use read count structural variant inferring logic when VAF = 1
purple-v2.33
- Raised QC segment fail to >220 unsupported segments
- Infer LOH in simple DUPs between 2 LOH regions
- Fixed HG38 regression bug
- Added new logic to structural variant recovery to create inferred variant if unable to find suitable candidate in file.
- Changed relative copy number tolerance when smoothing from fixed 10% to 0.12 + 0.8 / sqrt(min depth window count)
- Fixed bug in subclonal plot
purple-v2.32
Fix IndexOutOfBoundsException in clonality modelling
cobalt-v1.7
- Exit gracefully on exceptions
- Changed file names and headers for better consistency with other HMF tools
sv-linx-v1.2
DEV-759: refactor
sv-linx-v1.3: Linx:
- reportable fusions file context marked as 'Exon' since always reports the fused exon