search

kiharalab / Domain-PFP

Domain-PFP is a self-supervised method to predict protein functions from the domains
GNU General Public License v3.0
7 stars 1 forks source link

Request for terms used in CAFA3 dataset #2

Open wlin16 opened 4 months ago

wlin16 commented 4 months ago

Hi,

Thank you for contributing to the protein function prediction. I noticed that only the terms used in NetGO2.0 benchmark were provided (terms.pkl) while that file of the CAFA3 dataset was not given.

I tried to reproduce the number of labels used for each sub-ontology by various methods, such as only considering downstream GO term nodes. However, I cannot get the exact same number of labels used for each category presented in the supplementary part of the publication. Therefore, I am wondering if it is possible to provide the terms.pkl file used for CAFA3 dataset as well?

Best regards WL

nibtehaz commented 4 months ago

Hi @wlin16

Apologies for the delayed response, thank you for your interest in our Domain-PFP work.

For the CAFA3 evaluation we used the CAFA3 data and terms processed by DeepGoPlus. You can find them here https://github.com/bio-ontology-research-group/deepgoplus