handle overlapping mates in M2 isActive

[davidbenjamin]

@takutosato This change doesn't hurt sensitivity in our validations and made M2 25% faster. We were getting a lot of active regions based on single substitution errors in overlapping reads.

[ldgauthier]

Nice!

[codecov-io]

Codecov Report

Merging #5078 into master will increase coverage by 0.071%.
The diff coverage is 100%.

@@              Coverage Diff               @@
##             master     #5078       +/-   ##
==============================================
+ Coverage     86.39%   86.461%   +0.071%     
- Complexity    28832     29228      +396     
==============================================
  Files          1791      1791               
  Lines        133730    135234     +1504     
  Branches      14916     15271      +355     
==============================================
+ Hits         115529    116924     +1395     
- Misses        12798     12849       +51     
- Partials       5403      5461       +58

Impacted Files	Coverage Δ	Complexity Δ
...der/tools/walkers/mutect/M2ArgumentCollection.java	93.75% <100%> (+0.417%)	3 <0> (ø)	⬇️
...hellbender/tools/walkers/mutect/Mutect2Engine.java	91.613% <100%> (+0.222%)	59 <0> (+4)	⬆️
...walkers/genotyper/GenotypingGivenAllelesUtils.java	40.541% <0%> (-20.998%)	4% <0%> (+2%)
...r/tools/spark/sv/utils/GATKSVVCFUtilsUnitTest.java	91.379% <0%> (-1.644%)	18% <0%> (+5%)
...ithwaterman/SmithWatermanIntelAlignerUnitTest.java	60% <0%> (ø)	2% <0%> (ø)	⬇️
...entedCpxVariantSimpleVariantExtractorUnitTest.java	99.699% <0%> (+0.086%)	48% <0%> (+16%)	⬆️
...nder/tools/spark/sv/utils/SVVCFWriterUnitTest.java	97.917% <0%> (+0.091%)	3% <0%> (ø)	⬇️
...der/tools/spark/sv/discovery/SvDiscoveryUtils.java	44.898% <0%> (+0.161%)	19% <0%> (+6%)	⬆️
...te/hellbender/tools/spark/sv/discovery/SvType.java	77.215% <0%> (+0.404%)	21% <0%> (+5%)	⬆️
...nce/SegmentedCpxVariantSimpleVariantExtractor.java	94.54% <0%> (+0.514%)	103% <0%> (+37%)	⬆️
... and 13 more