Fixing empty group case
Fixing MD5s
First comments addressed
Added permutation test
Adding new RankSum to AS_RankSum
Speeding up permutation algorithm and updating MD5s
Missed a few tests
Addressing comments
Changing md5s
Intermediate commit for tests
Adding tests
Fixing tests after rebase
Fixing one MD5
Fixing documentation
Removing annotation from standard group
Adding documentation
ReadAdaptorTrimmer (unsound and untested)
BaseCoverageDistribution (redundant with DiagnoseTargets)
CoveredByNSamplesSites (redundant with DiagnoseTargets)
FindCoveredIntervals (redundant with DiagnoseTargets)
VariantValidationAssessor (has a scary TODO -- REWRITE THIS TO WORK WITH VARIANT CONTEXT comment and zero tests)
LiftOverVariants, FilterLiftedVariants and liftOverVCF.pl (in #1106) (use Picard liftover tool)
sortByRef.pl (use Picard SortVCF)
ListAnnotations (useless)
Also deleted the java archive from the private repository (old junk we never use)
Grouped default output annotations to keep them from getting dropped when -A is specified; addresses #918
Also refactored code shared by ExcessHet and InbreedingCoeff
Integration Tests
Updated test
Changed method
Minor changes
Changed whitespace
Fixed uncalled counts and 0 in R
Fixed ReadBackedPileUp
Removed imports and changed MD5
Fixed failing test
Adding vqslod color
Updating script to create KB
Fixing integration test now that the KB is bigger
Adressing comments
Updated other IntelliJ IDEA warnings in GATKBAMIndex.
Updated example .cram files to match versions generated by current GATK/HTSJDK.
Bumped HTSJDK and Picard to 1.139 releases.
Added support for using `-SNAPSHOT` of HTSJDK in the future.
This change doesn't affect the performance of the Indel Realigner at all (as per tests).
This is just a request from the Picard side (where further testing is happening).
Make MQ threshold a parameter (compare to M1 by setting to zero)
Add logic for multiple alternate alleles in tumor
Exclude MQ0 normal reads from normal LOD calculation
Fix path errors in Dream_Evaluations.md
Move M2 eval scripts out of walkers package so they run
CRAM now requires .bai index, just like BAM.
Test updates:
- Updated existing MD5s, as TLEN has changed.
- Tests multiple contigs.
- Tests several intervals per contig.
- Tests when `.cram.bai` is missing, even when `.cram.crai` is present.
Updated gatk docs for CRAM support, including:
- Arguments that work for both BAM and CRAM listed as such.
- Arguments that don't work for CRAM either explicitly say "BAM" or "doesn't work for CRAM".
- Instructions on how to recreate a `.cram.bai` using cramtools.
Cleaned up IntelliJ IDEA warnings regarding `Arrays.asList()` -> `Collections.singletonList()`.
Changed a division by -10.0 to a multiplication by -.1 in QualUtils (typically multiplication is faster than division).
Addresses performance issue #1081.
Add oxoG read count annotation and add as default annotation
Add ##SAMPLE VCF header line in accordance with TCGA VCF spec, specifying "File" line in sample header with BAM file name and "SampleName" with BAM sample name (Don't print sample file path if --no_cmdline_in_header is specified to help with test consistency)
Turn on active region assembly-based physical phasing for M2
Clean up M2-related annotations so UG doesn't crash if M2 annotations are called
added "str_contraction" artifact filter (improves specificity, especially in exomes)
refactored out VCF constants and added descriptions
added "artifact detection mode" for PON creation
added "str_contraction" artifact filter (improves specificity, especially in exomes)
added new dream evaulation markdown
added results for SMC 4
fixed up documentation, moved location to /dsde/working/mutect/dream_smc, and checked in scala script
added "artifact detection mode" for PON creation
added "str_contraction" artifact filter (improves specificity, especially in exomes)
fixed bug which would overwrite germline_risk filter errors
updated "how to" documents and records
fixed license text
thinned down FP regression test from 700 sites to 100. we have better ways (DREAM, NN) to check accuracy of the method and 100 is good enough to catch regressions
why oh why do the MD5-based unit tests produce different results on different machine architectures? I hate that :/
Thanks to GG, LDG and DR -- test should now produce the same results regardless of machine architecture
disabled downsampling... hopefully in the final attempt to make this work cross architecture!
enforced LOGLESS_CACHING... hopefully in the final final attempt to make this work cross architecture!
refactored out VCF constants and added descriptions
Geraldine Van der Auwera
Ron Levine
meganshand
Yossi Farjoun
Takuto Sato
Laura Gauthier
Eric Banks
Samuel Lee
Khalid Shakir
Bertrand Haas
vruano
Valentin Ruano Rubio
Joseph White
Kristian Cibulskis