I am using the tool "hap.py" from GA4GH for comparing variant callers (vcf files) and following the documentation "https://github.com/Illumina/hap.py/blob/master/doc/happy.md#happy-users-manual".
But, TRUTH.TOTAL get differ in each comparison cases. As truth variants are stable across all comparisons, the total truth variant count should be the same.
Can anyone explain it better for me?
•
written
by
Palani Kannan •
Yes, I would expect that TRUTH.TOTAL should be the same across all of the comparisons, provided the only thing that you are changing between comparisons is the call set. If you change other configuration (e.g. evaluation regions, normalization options, comparison engine etc) then the TRUTH.TOTAL may change. If you double check that and you still see differences, it may be worth filing an issue on the hap.py github page.
I have tried all those mentioned options. I couldn't solve the issue. I moved out of GA4GH to another tool for its complexity. Thanks for your suggestion.
Log in to answer this question.
