broadinstitute / gatk

Official code repository for GATK versions 4 and up
https://software.broadinstitute.org/gatk
Other
1.72k stars 594 forks source link

Eliminate redundant AnnotatedInterval class. #3884

Open samuelklee opened 7 years ago

samuelklee commented 5 years ago

Per discussions with @fleharty, we are looking to significantly revamp the automated somatic CNV evaluations in preparation for benchmarking the TH prototype. The existing evaluations use a few unsupported/experimental tools and idiosyncratic/redundant classes (e.g., the src/main/java/org/broadinstitute/hellbender/tools/copynumber/utils/annotatedinterval class this issue concerns), the functionality of which we can hopefully move to python-based validation code.

The aforementioned code was purposefully decoupled from supported CNV code, but since then it has been incorporated into Funcotator tools and ValidateBasicSomaticShortMutations, at least. @jonn-smith @davidbenjamin can we discuss a plan for cleaning this code up? Would it be easy to use an existing TSV/XSV class to handle the functionality needed for these tools?

@jonn-smith perhaps we should also discuss the plan for future FuncotateSegments development/integration with @fleharty.

davidbenjamin commented 5 years ago

Its use in ValidateBasicSomaticShortMutations seems limited to the integration test. Can I rewrite the test to do without AnnotatedInterval and call it a day?

samuelklee commented 5 years ago

That would be great, thanks @davidbenjamin!