The details of the preprocess of GraphDTI is provided here.This procedure is used specifically for the GraphDTI project (https://github.com/Guannan1900/GraphDTI).
- graph2vec generation: generate the Graph2vec features of the target proteins. Note that the Graph2vec features is used for GraphDTI project.
- graph2vec optimization: optimize the graph2vec features fot the target proteins.
- feature integration: integrate four types of features which are used for GraphDTI project.
- feature selection: select the optimal features for GraphDTI in order to mitigate the overfitting problem.
- clustering: design a cluster-based split protocol for cross-validation.
If you find this tool useful, please cite our paper :)