Extract and emit UniProt classes for proteins
Opened this issue · 1 comments
To help transition out of the REACTO convention, a low-hanging fruit is simply extracting the UniProt ID xrefs for proteins from the Reactome BioPAX and instantiate using these UniProt classes rather than the REACTO class.
Ex protein is in pathway R-HSA-163765 "ChREBP activates metabolic gene expression," reaction R-HSA-163666 having enabler protein MLX R-HSA-163695:
This change should switch from using class R-HSA-163695 to UniProtKB:Q9UH92.
This sounds right.
Can the protein's R-HSA-# ID be preserved as a cross-ref in the resulting GO-CAM instance? The rationale is that these identifiers for human proteins are true, well-behaved stable IDs, so will persist for as long as we have evidence that the protein (with its location and covalent modifications) really exists. If the evidence changes and we obsolete the protein, our history-tracking system should allow a user to see what, if anything, has replaced it.
Feature creep: this cross-rferencing will also be useful for the MOD projections. @ukemi