search

clingen-data-model / allele

Documentation for data model of ClinGen
10 stars 2 forks source link

Some amino acid change types are not disjoint #140

Closed cbizon closed 9 years ago

cbizon commented 9 years ago

The primary-amino-acid-change-type

http://datamodel.clinicalgenome.org/allele/implementation/value_set_list/primary_amino_acid_change_type.html

were initially supposed to be a disjoint set of annotations so that each amino acid allele could be assigned a single type algorithmically.

Ronak pointed out that frameshift can happen in conjunction with gains/losses of start/stops, we we need to decid what to do:

Stop_gained or frameshift? NP_000305.3:p.Thr319Terfs (http://www.ncbi.nlm.nih.gov/clinvar/variation/71118/)

start_lost or frameshift? NP_000050.2:p.Met1Ilefs: (http://www.ncbi.nlm.nih.gov/clinvar/variation/37872/)

cbizon commented 9 years ago

In both cases, I would call these frameshifts. It is true that there is also a start lost and a stop gained; I would put these in the auxialliary-amino-acid-change-type.

In other words, frameshift has priority over the other types, even if they are not disjoint.