SAGE VCF Header Update Definition Issue

qclayssen commented 1 week ago

The current VCF header definitions for the SAGE(V2021) fields are misaligned with their last version(V2024) data structures, causing validation errors.

Issue: The SAGE_SB field contains two float values separated by a comma (e.g., 0.5,0.457), but the header specifies Number=1, indicating only one value. This discrepancy results in validation errors such as "invalid float."

RESCUED_FILTERS_PENDING field may contain multiple comma-separated values (e.g., FILTER1,FILTER2), but the header specifies Number=1, which expects only one value.

Slack thread Branch

qclayssen commented 1 week ago

One of the problems could be retro compatibility. @scwatts

Do we want one version of Bolt to handle only one version of SAGE outputs?
Or should Bolt include a check for the SAGE version and dynamically apply the correct header definitions?

scwatts commented 1 week ago

I'd prefer a given version of bolt to support only a specific version of SAGE for simplicity - I don't think we really have a use case where we'd want backwards compatibility. We at least haven't need this previously and I'm not able to come up with a realistic scenario but let me know if you have some situation in mind.

It's reasonable to assume that if we wanted outputs with the most recent release of bolt we'd also want to use the most recent version of oncoanalyser, so I'd say we're safe to proceed without backwards compatibility.

We should take this approach done for PCGR and extend to SAGE to avoid the issues recently encountered.

scwatts / bolt

SAGE VCF Header Update Definition Issue #4