Figure 1

The systematic workflow of our methodological approach.
Drug targets (DTs) were obtained by collecting information from different pharmaceutical company websites in the different phases of the drug discovery process (in red, yellow and orange). The interaction network of DTs was then constructed by extracting the direct 1-step neighbors of the DT based on the i2d database (the blue nodes in the network). Following the integration of multiple and heterogeneous data types by using the MRAM method, the rules were induced to predict the potential DTs. Finally, We characterized the functionality of the potential DTs by testing over-represented Gene Ontology biological process terms and pathways.