Capture classification context reason

[theferrit32]

During the conditional processing when computing classification context, failures and successes of conditions can be tracked and provided as a reason string alongside the classification context annotation.

We should clarify how useful this would be and also what the contents of those strings should be.

For example a SOMATIC_CANCER annotation could be explained with the programmatically generated reason string Assertion had only somatic allele origins; had subjects on cancer gene(s) HRAS; did not have predicate risk factor

The strings had only somatic allele origins, had subjects on cancer gene(s) HRAS, did not have predicate risk factor can be recorded at each step. If it fails one of these conditions, an opposing reason string would be added to the list saying why it was not somatic cancer. In the case of GERMLINE_DISEASE, the reason string would contain the reasons it is not somatic cancer or pharmacogenomic, plus the reason(s) it is germline disease.

[larrybabb]

@dazzariti we should discuss the benefits of capturing the reasons for how each SCV is binned with Heidi. It may be useful if we question why something is in a given bin and it may be useful to future users that may be new to our dataset and might get immediate benefits from inspecting the reasons if desired.

[KelseaChang5]

December 7th 2021 Triage: Renamed from Clarify and implement classification context reason annotation during classification context computation to Capture classification context reason