This repository explores the correlation between gene expression, methylation and copy number with respect to the OS (Overall Survival) of a cancer patient, more specifically an ovarian cancer patient.
To explore this we use different types of Deep Learning model, and we represent the data with a graph structure.
More information about this project and its results are available in the following paper.
- GGNN: External repository from the following paper, A Gemetric informed Graph neural network based model for Cancer survival prediction.
- GNN_Classification: External repository from link.
- Preprocessing: preprocessing of input data and creation of relevant files used during training, in this folder we tested different methods to apply the preproces, a final version for the preprocessing can be observed in the "Training" folder.
- Training: actual training.
- requirements.txt: library requirements to run the repository (some might not be necessary, since they may have been used for testing some functionality not used in training).