DiMiMe model uses microbial information as a bridge to fuse disease and metabolite information through bipartite graph attention network, and finally realizes the association prediction of the disease-metabolite.

All node features and edge data of bipartite graph. Positive samples and negative samples of all kinds.

Result data used in experimental case analysis.

K-means method used in negative sampling.

Model code, initialization function and evaluation metrics function involved.

The model is tested to work under python3.6. The required dependencies versions are as follows:

torch==1.4.0+cu100
torch-cluster==1.5.2
torch-geometric==1.4.1
torch-scatter==2.0.3
torch-sparse==0.5.1
torch-spline-conv==1.2.0
torchvision==0.5.0+cu100
scikit-learn==0.21.3
pandas==1.1.5
numpy==1.16.0