Incorrect mapping of CHEK2

[jstone-dev]

In MaveDB, CHEK2 variants have been mapped to NP_001005735.1, which is a poor choice, rather than the MANE Select protein ID NP_009125.1. The cause appears to be the fact that UTA has an outdated MANE Select transcript version (NM_007194.3; NM_007194.4 is current). The mechanism by which this leads to poor transcript selection is discussed at https://github.com/bbi-lab/variant-annotation/blob/main/docs/map_variants.md#transcript-selection.

I'd suggest that this could be addressed by

• Updating UTA (if it now contains the current version)
• Adding an option to force transcript selection (as has been done in the standalone pipeline linked above)
• Or (perhaps) loosening the UTA query so that we find older versions of MANE transcripts. This approach might involve assumptions we don't want to make, so it should probably be accompanied by flagging the mapping somehow.