zjunlp / OntoProtein

[ICLR 2022] OntoProtein: Protein Pretraining With Gene Ontology Embedding
MIT License
141 stars 22 forks source link

Graph knowledge creation #19

Closed Amels404 closed 1 year ago

Amels404 commented 2 years ago

Hello, I have been looking at this repo for the last couple of days. I'm interested in graph knowledge generation from the go features. How did you manage to create the graph? Is there any snippet of code or detailed documentation? Thank you!

Alexzhuan commented 1 year ago

Hi,

We constructed the protein knowledge graph by associating Gene Ontology with the GO annotations for proteins.

We have provided the code of graph construction OntoProtein/gen_onto_protein_data.py, which could be used to generate a protein knowledge graph (saved as triplets, including the two types of triplet, Protein-GO triplet and GO-GO triplet).