Since the consequences are ranked via priority (not score) for rank models using VEP CSQ instead of most_severe_consequence, we should set the correct priority to match score. Here for example, three different terms with different scores have the same priority, meaning the highest scoring term will not be the one chosen, rather the one appearing first in the rank model file:
|
[[non_coding_transcript_exon_variant]] |
|
score = 3 |
|
priority = 2 |
|
string = 'non_coding_transcript_exon_variant' |
|
|
|
[[synonymous_variant]] |
|
score = 2 |
|
priority = 2 |
|
string = 'synonymous_variant' |
|
|
|
[[start_retained_variant]] |
|
score = 1 |
|
priority = 2 |
|
string = 'start_retained_variant' |
I can find no documentation for this, either in genmod or atlas and genmod does not complain when multiple entries have the same priority. It should clear and easy for a users who wants to add or update rank models how they work, not a mystery.
Acceptance Criteria
Notes
- Additional information.
- Dependencies.
- Related user stories.
Since the consequences are ranked via
priority(notscore) for rank models using VEPCSQinstead ofmost_severe_consequence, we should set the correctpriorityto matchscore. Here for example, three different terms with different scores have the same priority, meaning the highest scoring term will not be the one chosen, rather the one appearing first in the rank model file:reference-files/nallo/rank_model/grch38_rank_model_svs_-v1.0-.ini
Lines 372 to 385 in 72eff3c
I can find no documentation for this, either in genmod or atlas and genmod does not complain when multiple entries have the same priority. It should clear and easy for a users who wants to add or update rank models how they work, not a mystery.
Acceptance Criteria
Notes