DCF is an approach to detect underlying relationships between genes and diseases by seamlessly combining the deep architecture, SDAE model for auxiliary side information and the collaborative filter for the gene-disease association matrix.
DCF operates on large collections of heterogeneous data sets, such as gene expression, gene-gene interactions, gene orthology, OMIM pages and disease similarity networks.
This repository contains supplementary material for Deep collaborative filtering for prediction of disease genes.
- run run_DCF.m to get result of specific dataset and get results.
- run code novel_eval.m to visualize the result of novel disease evaluation. Results of fig 4(b) is given, you can modify it for other situations likewise.
-
You can download data from Inductive matrix completion for predicting gene-disease associations
-
for more, you can contact with me, some files is too large to upload.