RTXteam / RTX-KG2

Build system for the RTX-KG2 biomedical knowledge graph, part of the ARAX reasoning system (https://github.com/RTXTeam/RTX)
MIT License
39 stars 8 forks source link

More specific categories for Interleukin 2 nodes (from LOINC/UMLS) #277

Open amykglen opened 1 year ago

amykglen commented 1 year ago

these three nodes have a category of NamedThing in KG2.8.2pre, which prevents them from being merged with the other nodes representing Interleukin 2 that have Protein/Gene or SmallMolecule categories during the synonymizer build:

match (n) where n.id in [
    "LOINC:MTHU015779",
    "LOINC:LP16470-4",
    "MESH:D007376"
  ] return n.id, n.name, n.category, n.description
n.id n.name n.category n.description
"LOINC:MTHU015779" "Interleukin 2" "biolink:NamedThing" "UMLS Semantic Type: STY:T129; UMLS Semantic Type: STY:T116"
"LOINC:LP16470-4" "Interleukin 2" "biolink:NamedThing" "UMLS Semantic Type: STY:T129; UMLS Semantic Type: STY:T116"
"MESH:D007376" "Interleukin-2" "biolink:NamedThing" "A soluble substance elaborated by antigen- or mitogen-stimulated T-LYMPHOCYTES which induces DNA synthesis in naive lymphocytes.; UMLS Semantic Type: STY:T121; UMLS Semantic Type: STY:T129; UMLS Semantic Type: STY:T116"

https://arax.ncats.io/devLM/?term=LOINC:MTHU015779

would it be possible to provide more specific categories for these nodes during the KG2pre build?

ecwood commented 1 year ago

In the Biolink model, this is how those TUIs are categorized:

One of the ancestors of drug is chemical entity. polypeptide falls under biological entity, but chemical entity and biological entity only share named thing a common ancestor. Thus, the only common ancestor for all three TUIs is named thing. However, this only applies to the MESH term. I will look more into the other two.