generate example results for "creative/inferred" mode for relay #453
Assignees
Comments
|
Posting from Slack, David K. from Translator's ask:
EDIT: there are 9 diseases/phenotypes posted below as of 6/6/22 8:49PM PDT. Common:
Rare: |
Alzheimers (MONDO:0004975)Initial queryAssuming the query given will be... Query starting pointFuture Responses:
Other methods that didn't return results:
Note! If biolink version is updated, we may need to change the predicates used in the methods below. "Straightforward method"
"straightforward" queryMethod 1: Disease X -(has_phenotype)-> PhenotypicFeature -(treated_by)-> ChemicalEntity
"method 1" queryMethod 2: Disease X -(superclass_of)-> Disease --> Gene -(affected_by)-> ChemicalEntity
"method 2" queryMethod 3: Disease X --> SequenceVariant --> Gene -(affected_by)-> ChemicalEntity
"method 3" queryMethod 4: Disease X --> Gene -(affected_by)-> ChemicalEntity
"method 4" queryMethod 5: Disease X -(has_phenotype)-> PhenotypicFeature -(affected_by)-> Gene -(affected_by)-> ChemicalEntity
"method 5" queryMethod 6: Disease X -(caused_by)-> Gene -(affected_by)-> Gene -(affected_by)-> ChemicalEntity
"method 6" query |
neurofibromatosis, familial spinal (MONDO:0008078)Initial queryAssuming the query given will be... Query starting pointFuture Responses:
Other methods that didn't return results:
Dev notes:replace Alzheimer's node in the above query-templates for the methods, with this info:Method 5 issues: walking through the logsUsed the older format for the query (n0-n3, e1-e3) old method 5 queryFirst two hops only take ~ 1 min logs before the last hop
related logslogs showing how large the size is nowERROR 1: SCORING FAILSand the query is super long and then these logs look relevant ERROR 2: can't return the response due to string issue? looks like a TON of results too (326849) and took ~ 2 min to generate the logs during results-assembly |
noonan syndrome (MONDO:0018997)Initial queryAssuming the query given will be... Query starting pointFuture Responses:
Other methods that didn't return results:
Dev notes:replace Alzheimer's node in the above query-templates for the methods, with this info:Method 6 issues: walking through the logsVery similar to what happened above with neurofibromatosis in the above post queryFirst two hops only take ~ 1 min logs before the last hop
related logslogs showing how large the size is nowERROR 1: SCORING FAILSand the query is super long and then these logs look relevant ERROR 2: can't return the response due to string issue? looks like a TON of results too (289878) and took ~ 1 min to generate the logs during results-assembly |
acanthosis nigricans (MONDO:0007035)Initial queryAssuming the query given will be... Query starting pointFuture Responses:
Other methods that didn't return results:
Dev notes:replace Alzheimer's node in the above query-templates for the methods, with this info: |
Abnormal blood glucose concentration (HP:0011015 / Phenotype)Initial queryAssuming the query given will be...(assigning this ID as a Disease even though it's not...BTE will add PhenotypicFeature to its categories) Query starting pointFuture Responses:
Other methods that didn't return results:
Dev notes:replace Alzheimer's node in the above query-templates for the methods, with this info:(assigning this ID as a Disease even though it's not...BTE will add PhenotypicFeature to its categories) |
chronic obstructive pulmonary disease (COPD, MONDO:0005002)Initial queryAssuming the query given will be... Query starting pointFuture Responses:
Other methods that didn't return results:
Dev notes:replace Alzheimer's node in the above query-templates for the methods, with this info: |
cirrhosis of liver (MONDO:0005155)Initial queryAssuming the query given will be... Query starting pointFuture Responses:
Other methods that didn't return results:
Dev notes:replace Alzheimer's node in the above query-templates for the methods, with this info: |
Colon cancer (HP term that likely maps to disease in Node Normalizer) (HP:0003003)Initial queryAssuming the query given will be... Query starting pointFuture Responses:
Other methods that didn't return results:
Dev notes:replace Alzheimer's node in the above query-templates for the methods, with this info: |
Diarrhea (HP:0002014 / Phenotype)Initial queryAssuming the query given will be...(assigning this ID as a Disease even though it's not...BTE will add PhenotypicFeature to its categories) Query starting pointFuture Responses:
Other methods that didn't return results:
Dev notes:replace Alzheimer's node in the above query-templates for the methods, with this info:(assigning this ID as a Disease even though it's not...BTE will add PhenotypicFeature to its categories) |
|
Modified Method 1: add is_set to Pheno node
|
|
Modified Method 2 (set predicates between disease-gene, use more-specific predicate between gene-chem, add is_set on intermediate nodes): Of note:
Basically the same behavior:
|
|
Modified Method 3 (use more-specific predicate between gene-chem, add is_set on intermediate nodes): May not want to use this because we can't score any results (SequenceVariants don't have UMLS IDs)... Of note:
Still problematic:
Basically the same behavior:
|
|
Modified Method 4 (add predicates between disease-gene, use more-specific predicate between gene-chem, add is_set on intermediate nodes): Of note:
Still problematic:
Basically the same behavior:
|
|
Relay is over; closing in favor of #461 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
There is an issue to track the implementation of a "creative/inferred" mode. At the upcoming relay, there is a session on "creative mode" on Monday and Tuesday in which we would like to explore some preliminary results. Organizers would like us to come with some example output for two queries from the list of example queries. Specifically, organizers would like to see results for "Alzheimer disease MONDO:0004975" and "neurofibromatosis, familial spinal MONDO:0008078". In advance of our full technical implementation, we can show up with simply a list of ARAX links (each corresponding to one of the query templates in https://github.com/Carolina1396/drug_mechanisms_rare_diseases_BioThingsExplorer/tree/main/src/query_templates).
This is meant to be a very rough approximation of what will come out later, just so we have something to discuss at the relay. So if the plan described above needs to be modified for the sake of easy feasibility, no problem... Initially assigning to @colleenXu.
The text was updated successfully, but these errors were encountered: