Making OB filtering default to true

[LeeTL1220]

• M2 WDL will now default to having the orientation bias filter (FilterByOrientationBias) turned on. Closes Set default to use the FilterByOrientationBias tool in the WDL #5016
• Empty artifact modes parameter will no longer cause failure in FilterByOrientationBias. Instead the task FilterByOrientationBias will not run. Closes M2 WDL will fail FilterByOrientationBias if artifact_modes is defined as empty and run_orientation_bias_filter is true #5025

[codecov-io]

Codecov Report

Merging #5019 into master will increase coverage by 26.174%.
The diff coverage is n/a.

@@               Coverage Diff                @@
##              master     #5019        +/-   ##
================================================
+ Coverage     60.162%   86.335%   +26.174%     
- Complexity     12772     28501     +15729     
================================================
  Files           1095      1779       +684     
  Lines          64616    132167     +67551     
  Branches       10394     14721      +4327     
================================================
+ Hits           38874    114107     +75233     
+ Misses         21504     12750      -8754     
- Partials        4238      5310      +1072

Impacted Files	Coverage Δ	Complexity Δ
...stitute/hellbender/utils/nio/ExtremeReadsTest.java	6.122% <0%> (ø)	2% <0%> (?)
...dateBasicSomaticShortMutationsIntegrationTest.java	100% <0%> (ø)	5% <0%> (?)
...nder/engine/filters/MetricsReadFilterUnitTest.java	100% <0%> (ø)	9% <0%> (?)
...nv/GermlineCNVIntervalVariantComposerUnitTest.java	96.61% <0%> (ø)	10% <0%> (?)
...r/tools/walkers/bqsr/ApplyBQSRIntegrationTest.java	79.104% <0%> (ø)	17% <0%> (?)
.../activityprofile/ActivityProfileStateUnitTest.java	86.667% <0%> (ø)	13% <0%> (?)
...te/hellbender/tools/PrintReadsIntegrationTest.java	96.795% <0%> (ø)	26% <0%> (?)
...s/vqsr/VariantRecalibratorModelOutputUnitTest.java	92.92% <0%> (ø)	23% <0%> (?)
...specific/AlleleSpecificAnnotationDataUnitTest.java	86.667% <0%> (ø)	4% <0%> (?)
...opynumber/CollectAllelicCountsIntegrationTest.java	100% <0%> (ø)	4% <0%> (?)
... and 1267 more

[davidbenjamin]

It looks like the default (i.e. neither run_orientation_bias_filter nor artifact_modes specified) leads to a faulty command line of -AM [nothing] in the FilterByOrientationBias task. Perhaps you could do something like Array[String] artifact_modes_to_use = if defined(artifact_modes) artifact_modes else ["C/T", "G/T"].

[LeeTL1220]

I am curious why the WDL auto tests did not catch this..

[LeeTL1220]

@davidbenjamin Back at you....

[davidbenjamin]

I am curious why the WDL auto tests did not catch this..

@LeeTL1220 In scripts/m2_cromwell_tests/test_m2_wdl_multi.json we test with

"Mutect2_Multi.run_orientation_bias_filter": true`,
"Mutect2_Multi.artifact_modes": ["G/T", "C/T"],

whereas the probable bug I foresaw would only occur if these lines were omitted.

[davidbenjamin]

Looks good but a thought on solving #5025 in this PR.

[davidbenjamin]

Could you solve this via (I didn't test this syntax)

Boolean run_ob_filter = select_first([run_orientation_bias_filter, true]) && length(final_artifact_modes) > 0

[LeeTL1220]

That would solve it in the workflow. Not necessarily the task if used in another workflow.

Done.

[davidbenjamin]

Back to @LeeTL1220

[LeeTL1220]

@davidbenjamin Back to you.

[davidbenjamin]

Looks good.