GeDGC

GeDGC Model

Input and Output Directories

The datasets used by the program are in the data folder. The output files are under the result folder and separated by the name of the dataset. i.e the results of the Chen_high are in result/Chen_high.

File description

The file descriptions are as follows:

graph_embedding: The files implementing the SpectralNet
module/Multi_vae.py: The file for designing the multi-omics variational autoencoder
module/classifier.py: The file for designing the classifier and Gaussian mixture models
tools/load_data.py: File for data import and data normalization
tools/Form_data.py: File for calculating neighbor nodes
tools/utils.py: File for calculating evaluation metrics
tools/early_stopping.py: File for the early stopping implementation
pretrain.py: A pre-training file for the model
train.py: A training file for the model
run.py: The main file for execution

Requirements

python                        3.7.12
muon                          0.1.3
scanpy                        1.9.3
torch                         1.13.1
torchvision                   0.9.1+cu101

Experiments

The algorithm can be run as follows:

python run.py