Fixing what looks like a a bug in MendelianViolationEvaluator

[lbergelson]

• there was a duplicated condition which looks like a typo and causes nNoCall to be undercounted

@bbimber
mv.getFamilyCalledCount was used twice as the condition in this if statement, I'm pretty sure it was a typo. As far as I know you're the one using this code so I want to check with you before changing it.

[bbimber]

@lbergelson, thanks for checking. We do run this with VariantQC, but I dont think people (at least not me) use this particular Evaluator much.

Looking at the code, the intention seems to be to skip not-called sites. Therefore if the condition is changed to:

        if(mv.getFamilyCalledCount()>0 || mv.getFamilyLowQualsCount()>0 || mv.getFamilyNoCallCount())

then would that ever be false? It seems like the family either has a non-zero mv.getFamilyCalledCount() or a non-zero getFamilyNoCallCount(), doesnt it? If so, what's the point of even testing that condition? I dont know that I'm right about this, but if that is correct then the other route is to just drop the redundant "mv.getFamilyCalledCount()>0".

            if(mv.getFamilyCalledCount()>0 || mv.getFamilyLowQualsCount()>0 || mv.getFamilyCalledCount()>0){
                nVariants++;
                nFamCalled += mv.getFamilyCalledCount();
                nLowQual += mv.getFamilyLowQualsCount();
                nNoCall += mv.getFamilyNoCallCount();
                nVarFamCalled += mv.getVarFamilyCalledCount();
            }