Matrix Factorization-Based Prediction of Novel Drug Indications by Integrating Genomic Space

Figure 1

Strategy pipeline. Firstly, feature vectors of genes are extracted from gene interaction network. Then, feature vectors of the same rank are obtained for drugs and diseases from drug-gene interactions and disease-gene interactions, respectively. Next, a matrix factorization model is generated to reconstruct the known drug-disease associations. Finally, the estimated feature vectors of drugs and diseases are used to infer new drug-disease associations and predict novel drug indications.