Exploratory analysis on drug-target interaction to enhance prediction of interaction

Abstract

Studies on Drug-Target Interaction (DTI) to discover the hidden relationships between drugs and targets in drug-target interaction are vital for drug development. Many researches had incorporated multiple drug-drug and gene-gene similarity measures for drug-target prediction to improve the accuracy of prediction. Visualization can be an effective analysis tool on DTI study because it could reveal the hidden structure of the networks and strengthen human understanding, thus leading to new insights, new findings and possible prediction of the future. In this project, Cytoscape plug-in was developed to aid DTI network visualization and to perform exploratory analysis on DTI by constructing similarities network involving drugs interact with different proteins and targets interact with different drugs. Based on the visualization results, we proposed a method to evaluate the quality of different similarities network, and compare the prediction performance of interaction from integration of different similarities measures. The results have shown that higher similarity network quality resulted in better prediction performance on drug-target interaction.