Created "variantcalling" test group and split them off from the rest of the integration tests for runtime purposes

[jamesemery]

Fixes #4983

[codecov-io]

Codecov Report

Merging #4984 into master will decrease coverage by 0.014%.
The diff coverage is n/a.

@@               Coverage Diff               @@
##              master     #4984       +/-   ##
===============================================
- Coverage     80.787%   80.773%   -0.014%     
+ Complexity     17966     17950       -16     
===============================================
  Files           1095      1095               
  Lines          64619     64619               
  Branches       10400     10400               
===============================================
- Hits           52204     52195        -9     
- Misses          8389      8395        +6     
- Partials        4026      4029        +3

Impacted Files	Coverage Δ	Complexity Δ
...oadinstitute/hellbender/utils/pairhmm/PairHMM.java	74.286% <0%> (-3.571%)	24% <0%> (ø)
...hellbender/utils/pairhmm/VectorLoglessPairHMM.java	82.895% <0%> (-1.316%)	11% <0%> (ø)
...e/hellbender/engine/AbstractConcordanceWalker.java	87.952% <0%> (-1.205%)	13% <0%> (-1%)
...rs/annotator/allelespecific/AS_StrandBiasTest.java	85.366% <0%> (-0.813%)	43% <0%> (-1%)
...kers/variantutils/PosteriorProbabilitiesUtils.java	81.41% <0%> (-0.641%)	50% <0%> (-8%)

[droazen]

One comment to address, then this is good to merge 👍

[droazen]

I think that the variantcalling tests should not be excluded in the else clause here, since the intention of the else clause is to include all the "normal" (unit/integration) tests, I believe.