KGANCDA

KGANCDA is effective to predict associations between circRNA and cancer, which is based on knowledge graph attention network. In this paper,we propose a new computational method (KGANCDA) to predict circRNA-disease associations based on knowledge graph attention network.The circRNA-disease knowledge graphs are constructed by collecting multiple relationship data among circRNA, disease, miRNA and lncRNA. Then, the knowledge graph attention network is designed to obtain embeddings of each entity by distinguishing the importance of information from neighbors.Besides the low-order neighbor information, it can also capture high-order neighbor information from multi source associations,which alleviates the problem of data sparsity. Finally, the multi layer perceptron is applied to predict the affinity score of circRNA-disease associations based on the embeddings of circRNA and disease.

Author: Wei Lan. Yi Dong. Qingfeng Chen. Jin Liu. Yi Pan. Yi-Ping Phoebe Chen.

Paper: https://academic.oup.com/bib/article-abstract/23/1/bbab494/6447436

Citation

If you want to use our codes and datasets in your research, please cite:

@article{10.1093/bib/bbab494,
    author = {Lan, Wei and 
              Dong, Yi and 
              Chen, Qingfeng and 
              Zheng, Ruiqing and 
              Liu, Jin and 
              Pan, Yi and 
              Chen, Yi-Ping Phoebe},
    title = "{KGANCDA: predicting circRNA-disease associations based on knowledge graph attention network}",
    journal = {Briefings in Bioinformatics},
    volume = {23},
    number = {1},
    year = {2021},
    month = {12},
    issn = {1477-4054},
    doi = {10.1093/bib/bbab494},
    url = {https://doi.org/10.1093/bib/bbab494},
    note = {bbab494},
    eprint = {https://academic.oup.com/bib/article-pdf/23/1/bbab494/42229911/bbab494.pdf},
}

Environment Requirement

• tensorflow == 1.12.0
• numpy == 1.15.4
• scipy == 1.1.0
• sklearn == 0.20.0

Dataset

Dataset 1 (Cancer)

including those data files:

• cancer_list.csv record cancer name and id.
• circRNA_list.csv record circRNA name and id.
• lncRNA_list.csv record lncRNA name and id.
• miRNA_list.csv record miRNA name and id.
• circRNA-cancer.csv record associations between circRNAs and cancers.
• circRNA-miRNA.csv record associations between circRNAs and miRNAs.
• lncRNA-cancer.txt record associations between lncRNAs and cancers.
• lncRNA-miRNA.txt record associations between lncRNAs and miRNAs.
• miRNA-cancer.csv record associations between miRNAs and cancers.

Dataset 2 (Non-cancer)

including those data files:

• circ_list.csv record circRNA name and id.
• dis_list.csv record disease name and id.
• lnc_list.csv record lncRNA name and id.
• mir_list.csv record miRNA name and id.
• circrna-disease.txt record associations between circRNAs and diseases.
• circrna-mirna.txt record associations between circRNAs and miRNAs.
• lncrna-disease.txt record associations between lncRNAs and diseases.
• lncrna-mirna.txt record associations between lncRNAs and miRNAs.
• mirna-disease.txt record associations between miRNAs and diseases.

Model

• KGANCDA.py: the core model proposed in the paper.
• main.py: the main program in the project. Run main.py to generate the embeddings of all nodes in the network
• ExtractFeature.py: After obtaining the embedding of all nodes, run this file to validate the effect of the model in 5-fold cross validation.

Compare_models

• There are 6 the state-of-the-art models including: CD-LNLP, DMFCDA, DWNN-RLS, GCNCDA, KATZHCDA, RWR, which are compared under the same experiment settings.

Question

• If you have any problems or find mistakes in this code, please contact with me: Yi Dong: dongyi6463@163.com