Skip to content

Known bad alignment - #828

Merged
Peter-J-Freeman merged 5 commits into
developfrom
known_bad_alignment
Jun 10, 2026
Merged

Peter-J-Freeman merged 5 commits into
developfrom
known_bad_alignment

Conversation

@Peter-J-Freeman

Copy link
Copy Markdown
Collaborator

No description provided.

Old version does not discriminate between raw and all in select_transcripts field. This does not match VV anf VF processing and leads to slower responses when only latest transcript versions are required. Now, raw returns all transcripts at all versions and all returns all transcripts at latest versions. Closes issue #804
Certain alignments are too bad to justify keeping. The list in the file transcript_map_data.py can be used to filter mapping options, even if VVTA is updated and new rogue alignments are identified.
Comment thread VariantValidator/modules/gene2transcripts.py Outdated
@John-F-Wagstaff

Copy link
Copy Markdown
Collaborator

Looks OK, just a question on the Suffixes, and a possible simplification. I would just hit approve but I think that the suffix issue needs thinking on at least first.

I do wonder about doing a latest transcripts only materialised view for VVTA, but how to deal with the differing (suffixed) Ensembl data is a fun question, and we would then need to plumb it in correctly later in multiple places.

@Peter-J-Freeman

Copy link
Copy Markdown
Collaborator Author

will fix this up, but I have a feeling that there are other instances in VV and VF where we need to fix too

@John-F-Wagstaff

Copy link
Copy Markdown
Collaborator

I believe that the intention was to just not use them for normal output, mostly, but to keep them to have them to map from if user input matched, with appropriate warnings, if we wanted to handle this later. But we never did get to being able to handle that. Not only as the ID for the user input would not match the ID for the ensemble data, so in that case so we would have to check specifically for this issue, but also because we would then have to handle the possible inputs interpretations for GRCh37/38 and make sure they pass through the rest of the code properly.

As such we probably just need to make sure nothing is crashing, for now, since this would be bonus feature development, not bug-fixing.

I think the intention was that if we got around to this to find both suffixed and resubmit the 2 suffixed with warnings?, e.g "TranscriptVersionWarning: Two versions of this ENSEMBL transcript exist with different sequence but the same ID and version number, one for GRCh38 and one for GRCh37. Neiter of these should be used in publications, but both versions have been resubmitted. Please check the output for possible discrepancies and select an appropriate new variant definition, you may need to use a RefSeq transcript instead." but only for inputs that used an ENS id that matched the transcript, not for genomic or gene id based queries)

Ensembl transcripts with the same
ID but requiring genome build
info require additional handling
@Peter-J-Freeman

Copy link
Copy Markdown
Collaborator Author

hope this push maintains the required format. We will add the other stuff to the ever expandin to do list :)

Comment thread VariantValidator/modules/gene2transcripts.py Outdated
Comment thread VariantValidator/modules/gene2transcripts.py Outdated
hdp.get_tx_for_gene has specific structures that need
to be filtered out. This can be plugged into all other
uses of hdp.get_tx_for_gene and adapted for other
similar lists from hdp
Allows the API to optionally set the shorthand vcf
format with a mod to the interface. The gui will keep
the toggle switched on i.e. True. Also, range increased
to 500 to allow capture or pretty much all vars in dbSNP

@John-F-Wagstaff John-F-Wagstaff left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good now if you want to push it.

@Peter-J-Freeman
Peter-J-Freeman merged commit bfe2bfd into develop Jun 10, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants