search

geneontology / api-gorest-2023

Repository created by AWS CodeStar.
0 stars 0 forks source link

Queries for causal models should detect metabolic pathways #3

Open dustine32 opened 4 months ago

dustine32 commented 4 months ago

The queries used to find causal models currently look for 2 or more consecutive causal relations between acivities. Metabolic pathway GO-CAMs are connected via this pattern:

(MF1) -has_output-> (some_chemical) -has_input-> (MF2)

An example of these metabolic pathway models is Yeast ergosterol. image

To support display of these models on various sites, we should add this has_output/has_input pattern as criteria to the following queries: https://github.com/geneontology/api-gorest-2023/blob/958712f53db0e72544820795f6a26f388675c8ad/queries/sparql-models.js#L322 https://github.com/geneontology/api-gorest-2023/blob/958712f53db0e72544820795f6a26f388675c8ad/queries/sparql-gp.js#L67

I started working on a SPARQL query to find these a while back: https://gist.github.com/dustine32/6f76e6f7e34fbf1f0476ebe888906dbd

We've talked about doing this before but I can't find a ticket for it so, tagging @thomaspd @kltm @vanaukenk if they have any more info on prioritizing this.

thomaspd commented 4 months ago

The yeast ergosterol pathway above shows up on the GO-CAM site, which I think is based on a SPARQL query, so is the SPARQL query already available?

dustine32 commented 4 months ago

Great question, @thomaspd! And it points to this model maybe not being the best example as it is already considered a causal model using the current "2 or more consecutive causal relations between activities" criteria through this chain of activities: image (Note: The blank label box is a UniProt ID A0A5G2RLK9 that doesn't resolve in NEO) This is why the model is displayed on the GO-CAM site despite primarily being a pathway connected by inputs and outputs.