PFPmatrix- use different bits for newly added dbnsfp scores

[nakib103]

We added AlphaMissense and ESM1b scores in replacement of MetaLR and MutationAssessor in #1176

Alphamissense takes 3 bit prediction instead of 2-bit as before. This requires updates to all the scores in the protein function tables for all the species we have data for.

In this PR we change other tools to use 2-bit as before, only the new tools, dbnsfp_alphamissense and dbnsfp_esm1b, that are added uses 3 bits prediction.

Test

Test using transcript ENST00000311351

sandbox urls -

• variation page - http://wp-np2-35.ebi.ac.uk:7070/Homo_sapiens/Variation/Mappings?db=core;g=ENSG00000175324;r=8:38163339-38176493;t=ENST00000311351;v=rs148018686;vdb=variation;vf=545300927#ENST00000520755_545300927_G_tablePanel
• variant table in transcript - http://wp-np2-35.ebi.ac.uk:7070/Homo_sapiens/Transcript/Variation_Transcript/Table?db=core;g=ENSG00000175324;p=ENSP00000310596;r=8:38163339-38176493;t=ENST00000311351
• variants in protein - http://wp-np2-35.ebi.ac.uk:7070/Homo_sapiens/Transcript/ProtVariations?db=core;g=ENSG00000175324;r=8:38163339-38176493;t=ENST00000311351

dbNSFP Alphamissense+ESM1b -
compare same scores as before

SIFT/Polyphen/CADD/REVEL
Check with current website, e.g -
https://www.ensembl.org/Homo_sapiens/Transcript/Variation_Transcript/Table?db=core;g=ENSG00000175324;p=ENSP00000310596;r=8:38163339-38176493;t=ENST00000311351

[jamie-m-a]

Thanks for reverting / adding this functionality