Network pharmacology to aid the drug discovery process

The drug discovery process is expensive and time consuming. Hundreds of drugs enter the research and development pipeline only to be dismissed in late phase due to some risky side effect or lack of efficacy in human subjects. Using existing knowledge of approved drugs we can compare experimental drugs' expression profiles and chemical structures to predict their mechanism of action, filtering out the drugs that will not survive the development process saving time and money. We can also use these approaches in combination with clinical data to repurpose and combine existing drugs for improved therapeutic use. Here we discuss many of the current approaches in network pharmacology which can aid in the drug discovery process. First, we describe the fundamental data that forms the basis of these approaches and investigate where we can find this data. Next, we present how to use different data types incorporating network approaches to model drug effects, including various tools and algorithms developed for this purpose. Finally, we present a global overview of how to apply all of these techniques for accurate side effect and new indication predictions.