Only one Revel score reported for overlapping genes

[dnil]

Thank you for a very useful service!

We noted only one Revel score reported for a variant hitting overlapping genes, and as it happened the lowest score was reported.

https://spliceailookup.broadinstitute.org/#variant=11-61727368-C-A&hg=37&bc=basic&distance=500&mask=1&ra=0

Selected VEP entries for reference (Revel score highlighted)

Something peculiar to this one, or is it a general issue with overlapping transcripts/genes?

[bw2]

Thanks for reporting the issue.

For variants that are absent from gnomAD, SpliceAI-lookup relies on the myvariant.info API
which returns only the 0.565 score for this variant:

I just emailed the myvariant.info API developers about this.

I also checked the Ensembl VEP API, but it does not return REVEL scores:
https://grch37.rest.ensembl.org/vep/human/hgvs/11:g.61727368C%3EA?content-type=application/json