Closed jsstevenson closed 2 weeks ago
Haven't tested, but adding alt_ac = mane_transcripts[0]["GRCh38_chr"]
here returns a response.
We need to update the genomic_pos
to GRCh38
@jarbesfeld if GRCh38 assembly fails, do we need to revert to GRCh37? I realized that I've been having the mapper fail if we're unable to liftover to GRCh38.
@korikuzma I think it makes sense to leave as is and have the mapper fail if liftover to GRCh38 is unsuccessful.
@jarbesfeld okay, we can always add a new issue if it’s needed in the future. I’ll update the docs later
Closed by #356.
Describe the bug
When calling
genomic_to_exon_coords
with a start position on an exon and an ending position in intronic space (or vice versa), an older accession, andnearest_transcript_junction=True
, different procedures (liftover etc) are performed on the two positions. This results in an error about the chromosome accessions of the start and end coordinates being out of joint.Steps to reproduce
Expected behavior
n/a
Current behavior
Gives a warning (really an error):
Possible reason(s)
In the private method
_genomic_to_transcript_exon_coordinate
, there's a logic branch forget_nearest_transcript_junction
, that ends with anif
check for whether the position is on an exon. If it's not, then it's handled there and returned, but if it is, it leaves this section of the code entirely and has additional processing performed in the rest of the method. This appears to result in one accession being swapped in for a newer one. Not sure exactly where the bug is or what the correct response is yet.Suggested fix
No response
Branch, commit, and/or version
main
Screenshots
No response
Environment details
n/a
Additional details
No response
Contribution
Yes, I can create a PR for this fix.