search

broadinstitute / seqr

web-based analysis tool for rare disease genomics
GNU Affero General Public License v3.0
176 stars 88 forks source link

adding pext scores as an in silico score in seqr #924

Open anneodonnell opened 5 years ago

anneodonnell commented 5 years ago

pext scores are Beryl's transcript aware expression score. I was hoping it could be added to the in silico list. We use it in the opposite way from most scores - we use the score to downweight variants but it can't really upweight a variant - as it identifies variants that are in regions of a gene that don't appear to be expressed.

Recommended color scheme 0-0.2 color, 0.21-1 gray we want the color to catch people's attention but not be red, yellow or green. What would you think about having these be purple (or whatever color that could catch attention without screaming look at me).

I'm not totally sure where the pext score data lives but it's in the gnomAD browser. Either Beryl, Nick or Matt should be able to help.

hanars commented 3 years ago

https://github.com/broadinstitute/hail-elasticsearch-pipelines/issues/236 for tracking pipline work

hanars commented 2 years ago

From @mike-w-wilson : We do not have pext scores for 38 so this will only be implemented in 37

@anneodonnell do we still want to add these? All the CMG data is 38 now

mike-w-wilson commented 2 years ago

With gnomAD v4, pext will be implemented on 38 but it will be on a different version of GENCODE than what seqr currently uses. Hail defaults to vep95 which we use in seqr... gnomAD and the new GTEx will be on 101 or something newer. Using the 38 pext scores will require us to update our vep config in the pipeline once this happens. I'll move this issue to the icebox in the pipeline workspace if we are waiting on gnomAD/GTEx.

lynnpais commented 4 months ago

Per Anne, hold off until we have GRCh38 generated pext scores - hopefully coming in a few months