README

"[Adaptive Multi-view Graph Convolutional Networks for Protein Function Prediction]"

Environment Settings

• python == 3.7
• Pytorch == 1.1.0
• Numpy == 1.16.2
• SciPy == 1.4.1
• pytorch_pretrained_bert == 0.6.2
• transformers == 2.8.0
• scikit-learn == 0.22.2
• fair-esm ==0.4.0

Usage

CUDA_VISIBLE_DEVICES=1 nohup python model_main.py 

Data

Link

• Protein sequence: (http://www.uniprot.org/)
• Gene Ontology: (http://geneontology.org/docs/download-ontology/)
• Protein functional annotation: (http://geneontology.org/docs/download-go-annotations/)
• **├─Homo sapiens (or Human)
• **├─Saccharomyces cerevisiae (or Yeast)

Usage

Please first unzip the data folders and then use.

Yeast/
├─Yeast_new.txt: Yeast protein sequence.    
├─Yeast_Linsim.mat: taxonomic similarity and direct acyclic graph of GO term.  
├─Yeast_Scores.mat: protein sequence scores based on the BLAST method.  
└─YeastNGOA_R.mat: Yeast protein functional annotations