search

dms-vep / dms-vep-pipeline

Pipeline for analyzing deep mutational scanning (DMS) of viral entry proteins (VEPs)
Other
2 stars 0 forks source link

pre-built functional scores #123

Closed jbloom closed 1 year ago

jbloom commented 1 year ago

@Bernadetadad, I'm just raising this issue now so we can discuss before I make changes.

Right now the functional scores for each repo are computed just from the no-antibody controls in that repo. This works well under the model where we keep a single running repo of everything. It doesn't work so well when we are having to branch off a repo for each study, as then sometimes a single antibody repo (like the 12-19 one) only has a few no-antibody controls and so is using different and probably less accurate functional scores than estimated using more replicates.

How do we fix this? Do we allow an option to access an external pre-computed set of functional scores from a larger experiment?

Bernadetadad commented 1 year ago

That's a good point. of course I could just copy all vsvg/no_antibody samples to each new repo split. But if we decide to allow access to external repo then we should maybe do so for codon-variant table as well.

jbloom commented 1 year ago

Yes, re-use of existing codon-variant tables is already enabled in #121.

I will work on allowing similar for functional scores.

jbloom commented 1 year ago

Here is an example of a repo that uses the pre-built mutation effects via prebuilt_muteffects: https://github.com/dms-vep/SARS-CoV-2_Omicron_BA.1_spike_DMS_12-19/