-
Notifications
You must be signed in to change notification settings - Fork 151
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
High memory usage due to 368 BRCA1 RefSeq transcripts - Transcript blocklist / allowlists or option for max transcripts #1732
Comments
Hi @davmlaw, Sorry for the delay in response and thanks for pointing us to this interesting case. The solution you mentioned is not achievable with the current vep options. But we can add a option such as This will need changing how we filter affected transcripts for Best regards, |
I aksed RefSeq how on 17th of March 2024 one of the most intensely studied genes of all time went from 6 to 368 isoforms overnight, without them being reported in other species, and only citing references from the 1990s (latest being paper from 2003) and they said:
|
Another option would be transcript allow/blocklists? Presumably that would allow quick rejects before all the calculations are done? This could also be generally useful as people often have their choice of transcript per gene, that isn't always the MANE one |
The current GRCh38 RefSeq annotation (GCF_000001405.40-RS_2023_10) has 368 transcripts for BRCA1 - up from 6 the previous version.
I can't see anything obviously wrong with the new sequences but I am going to take it up with RefSeq and see what's going on.
In the meantime, this is causing a lot of trouble for VEP annotation runs as this is a commonly sequenced disease gene, and in my pipeline a single variant eg 17:43093010 G/A has a CSQ size of 475294 bytes
Perl was killed OutOfMemory after exceeding 16G of RAM as the only job. I had to drop down the buffer size from 5000 to 2000 top stop it being killed. It hasn't crashed yet but the run time and amount of annotations are extremely large
An option here of eg "--max_transcripts=50" would be great, but the trick would be not to have to calculate them all
If pick is based on something to do with the transcript (MANE, length etc) that's easy, but if using "most damaging", you'd ideally want to calculate just the damaging effects, then sort, then only proceed with further annotation (eg custom etc) on the ones you keep
The text was updated successfully, but these errors were encountered: