Description

Drug-protein networks have received considerable attention in recent years, given their relevance to pharmaceutical innovation and the production of new drugs. Many in silico approaches to predicting interactions based on network analysis and machine learning have been proposed with the aim of identifying new pharmacological interactions. However, there are still few tools that make online access available to the prediction platform for the scientific community. In this project, we propose the development of an online platform for predicting drug-protein interactions, through the integration of heterogeneous information sources. A differential of the proposed platform in relation to the existing ones lies in the fact that the prediction method to be used is able to select the most relevant sources of information to the problem and to make an optimal combination of data in order to produce better predictions . The proposed method is based on alternating optimization of predictor parameters and data combination coefficients. The developed platform will be available for free through a web service

Website

Tool Website