A NOVEL METHOD FOR DRUG TARGET INTERACTION PREDICTION

Prediction of drug-target interactions can be done by experimentally and it is very expensive and time consuming. Therefore, there is a continuous demand for computational techniques for more accurate predictions of interactions. Heterogeneous data sources are used to predict drug-target interactions by different approaches. Many algorithms have been formulated to infer novel drug - target interactions. But many of these algorithms had difficulty in predicting drug-target interactions involving new drugs or targets for which there are no known interactions. Also, most of the present computational techniques focus on a single representation of drugs or proteins. It has been shown that integrating multi-view representations of drugs and proteins can strengthen the prediction ability. For example, a drug can be represented by its chemical structure, or by its chemical response in different cells. A protein can be represented by its sequence, or by its gene expression values in different cells. The docking of drugs and proteins based on their structure can be considered as one view (structural view), and the chemical performance of them based on protein sequence and drug chemical structure can be considered as another view (chemical view). Thus many views of drugs and targets can be used for drug target interaction prediction: structural view and chemical view. In this work, a single-view approach of GRMF( Graph Regularised Matrix Factorization) , a matrix factorization method that use graph regularization, and then extend it to a multi-view approach , which could integrate both views using co regularized spectral clustering and multi-view low rank embedding methods . A pre-processing step is also used to enhance predictions in the “new drug" and “new target" cases by adding edges with intermediate interaction likelihood scores. The experiments show that this method predicted the drug-target interactions reasonably well.