Closed ValWood closed 9 months ago
first part of this, getting the numbers, I put at high priority so we can see what we have
It's easy to query Chado to get all the legacy interactions, but harder to query the genotype-genotype interactions.
Are these the numbers you were after?:
Old style genetic interactions:
interaction_type | evidence | count |
---|---|---|
interacts_genetically | Dosage Growth Defect | 98 |
interacts_genetically | Dosage Lethality | 94 |
interacts_genetically | Dosage Rescue | 1164 |
interacts_genetically | Negative Genetic | 67112 |
interacts_genetically | Phenotypic Enhancement | 1586 |
interacts_genetically | Phenotypic Suppression | 1890 |
interacts_genetically | Positive Genetic | 29482 |
interacts_genetically | Synthetic Growth Defect | 5422 |
interacts_genetically | Synthetic Lethality | 3241 |
interacts_genetically | Synthetic Rescue | 1612 |
Physical interactions:
interaction_type | evidence | count |
---|---|---|
interacts_physically | Affinity Capture-Luminescence | 3 |
interacts_physically | Affinity Capture-MS | 8618 |
interacts_physically | Affinity Capture-RNA | 456 |
interacts_physically | Affinity Capture-Western | 2637 |
interacts_physically | Biochemical Activity | 586 |
interacts_physically | Co-crystal Structure | 190 |
interacts_physically | Co-fractionation | 143 |
interacts_physically | Co-localization | 35 |
interacts_physically | Co-purification | 880 |
interacts_physically | FRET | 9 |
interacts_physically | Far Western | 35 |
interacts_physically | PCA | 1401 |
interacts_physically | Protein-RNA | 23 |
interacts_physically | Protein-peptide | 84 |
interacts_physically | Proximity Label-MS | 303 |
interacts_physically | Reconstituted Complex | 1559 |
interacts_physically | Two-hybrid | 3762 |
Note to self:
select p.value as evidence, rel_t.name as interaction_type, count(rel.feature_relationship_id) from feature_relationship rel join feature_relationshipprop p on rel.feature_relationship_id = p.feature_relationship_id join cvterm pt on pt.cvterm_id = p.type_id join cvterm rel_t on rel.type_id = rel_t.cvterm_id where rel_t.name like '%interact%' and pt.name = 'evidence' group by p.value, rel_t.name order by rel_t.name, p.value;
OK, thanks. I wondered if we used them all consistently thats all, it seems all are fairly well used. I'm going to close this ticket, until prompted by BIOGRID, because I can't remember what the pending changes were, maybe they never happened?
BIOGRID are changing the way that genetic interactions are described
I'm trying to make a plan (no hurry to implement, might take 2 years, but there might be small changes we could make to curation procedures now, to makes the eventual transition easier).
I will address the existing "types" one by one on the curation tracker first and curators can make a plan.
So for now I have a question. Could you let me know how many annotations we have to each GI term (synthetic lethal etc) (Is there anywhere I can look this up? I think there might be but I can't remember where it is)