This repository has been archived by the owner on Aug 23, 2024. It is now read-only.
Releases: cancerit/dockstore-cgpwgs
Releases · cancerit/dockstore-cgpwgs
2.1.0 - Performance, security and scientific correction
- Update base image to dockstore-cgpwxs:3.1.6 (from 3.0.3)
- Includes, memory footprint reduction for Caveman flagging
- Significant speedup in indel input generation
- Parallel flagging for SNV
- Fragment based SNV/Indel calling
- Upgrade allele counter and brass
- Reduce number of layers in final image.
- Remove some legacy env vars from wrappers
- Battenberg removed
- Was only included to do base allele counts not complete analysis
- Underlying data files are not supported for new builds or other species
- Contact original authors for ongoing support
- Wrapper script didn't explicitly specify sex SNP loci file (ASCAT)
- no impact for GRCh37 other species/build would have used wrong sex chromosome SNPs
- CaVEMan step wasn't being passed normal contamination value, scientific impact:
- previously
caveman.pl
default of0.1
was applied - this will now honor the value from ASCAT (
1-$rho
)
- previously
1.1.5 - scientific impact (backport)
Normal contamination derived from rho
value of ASCAT was not being propagated to CaVEMan. Default value of 0.1
has been used up to this point.
This is a backport to the legacy 1.1.x series of releases.
This fix will be added to the next 2.1.x release.
1.1.4 - More Pindel CPU control + fix for ASCAT -force
This is a back-port for problematic samples in the PanProstate analysis.
- Provide parameter to allow user to lower max pindel threads for analysis proving to have huge memory spikes.
- For CWL optionally provide an integer for
pindelcpu
in yourjson
- For singularity provide
-pindelcpu
to theds-wrapper.pl
script
- For CWL optionally provide an integer for
- Add
-force
toascat.pl
execution, seems to have been lost along the way
2.0.0 - support for both singularity and GRCh38
2.0.0
Major work to enable support for both singularity and GRCh38. Note that there may be an additional release to handle an efficiency issue in cgpPindel under GRCh38. Reference bundles for GRCh38 are not yet public.
- Full support for BAM+CSI and CRAM+CRAI as inputs
- Reorganisation to allow access of other tools or necessary bindings
- See dockstore-cgpmap 3.0.0 - 3.0.4 (primarily adds mismatchQc and threads to some tools)
- See dockstore-cgpwxs 3.0.0 - 3.0.3 (fragment based analysis)
- You will need to update the reference pack to include the new flagging rules, see example
json
files
- You will need to update the reference pack to include the new flagging rules, see example
- Using build stages to shrink images.
- remove legacy PRE/POST-EXEC from cgpbox days
- resolves #17, #18, #19, #20, #21, #24
- First official support for Singularity (extra variables and documentation)
1.1.3 - Pindel max CPU
- Limit max threads for pindel to prevent memory explosion on noisy datasets
- Minimal effect on run time as caveman processing is the limiting factor here
This is a back-port building based on the 1.1.2 image due to R library changes that would be too large a change to implement in a hotfix.
1.1.0 - Result reproducibility, I/O and runtime improvements
BRASS updates
- Significantly reduces the run time of brass_group phase
- see v6.0.0
- Requires new version of the CNV_SV reference archive
- Ensure stable result between replicate runs:
- see v6.0.2
Other dependency updates
- Based on dockstore-cgpwxs:2.1.1
- Update caveman, pindel and vagrent to improve reproducibility.
- Reduced I/O in caveman.
- Updates to alleleCount:v3.3.0 to improve access pattens for dense loci counting, resulting upgrades to:
Workflow updates:
- Reorganisation of some processes to reduce run-time:
- Changes to alleleCount have vastly improved I/O pattern and reduced runtime for
ascat and IMPUTE-allele counts (pre-Battenberg). - Reduced number of split files for IMPUTE-allele counts as reduced run time negates need for large spread
- Reduced alleleCount runtim allowed reorganisation of processes reducing runtime on example data from 23h to 19.5h (28cpu).
- Changes to alleleCount have vastly improved I/O pattern and reduced runtime for
- Added
cavereads
to example json specifying a higher value than default to reduce jobs required.
1.0.7 - BRASS upgraded to v5.4.0
1.0.7
- Update BRASS to v5.4.0
- Specifically fixes a vector merge bug which produces warnings (but no apparent change to results).
1.0.6 - cgpPindel bugfix + BAM header parsing
1.0.5 - stability and extra param
- Incorporates changes to reduce CPU oversubscription
- Exposes param for modifying caveman split sizes
More changes to improve stability and provide more info on environment
Improves debug and ability to reduce oversubscription