gnomAD heterozygous variants

[gurdeep330]

hey @tschmenger

We have new ML models now. (There is a long story behind that I must explain to you when you have >10 mins to spare).

Anyway, I need to regenerate the predictions of the heterozygous gnomAD variants that you looked at before. I will run the ML models for every variant but that can take days. So if you can already provide the list you looked at (mechismo-styled format- gene/mutation or uniprotAcc/mutation), I can run them ASAP, and you can update the MS (if the need be)

thanks

[tschmenger]

These were the variants I looked at and described in the manuscript:

P06493/V124G P46734/A84T Q02156/R510Q Q06418/V669L Q13470/A299D Q13557/D167E Q16832/T692N Q96RU7/S146N Q9UIK4/R68Q

The file is also found here

[gurdeep330]

@tschmenger Great, thanks!

Here are the results: gnomAD_het/

Give preference to AIvLD and RvN predictions for now. Hopefully, you wouldn't have to make many changes in the MS.

[tschmenger]

I think with the new predictors we have to revisit this. I assume no changes will be necessary for the manuscript, but currently the columns present in the file do not reflect the latest (and hopefully final) selection of predictors, iirc.

[gurdeep330]

@tschmenger, I redid the calculations: gnomAD_het/

[tschmenger]

Thanks for those! I will eventually need to compare all possible kinase mutations vs. gnomad for the table - or shall we omit those and only report the 9 above? All the others are present in the supplementary table, but are filtered out and not visible initially.

[gurdeep330]

Let's update the MS with these 9 for now and then have a look at the others on Monday when predictions for all kinases are ready.