Identifying pharmaceutical technology opportunities from the perspective of adverse drug reactions: Machine learning in multilayer networks

Identifying pharmaceutical technology opportunities can help guide drug innovation effectively and reduce the failure rate of pharmaceutical research and development (R&D) activities. Multiple factors are neatly associated with pharmaceutical R&D activities. Adverse drug reaction (ADR) is one of the main reasons for drug R&D failure. However, the current technology opportunities identification (TOI) approaches seldom try to identify pharmaceutical-specific technology opportunities from the perspective of ADR. Except that, pharmaceutical R&D activities are a complex process involving the interaction of disease, drug, and gene. Nevertheless, pharmaceutical TOI research often overlooks the relationships among various diseases, ADRs, drugs, and genes. In response, this article proposed a domain-specific multilayer network framework for pharmaceutical TOI from the standpoint of ADRs. Compared with previous research, there are three extensions. The first extension is identifying pharmaceutical TOI from the ADR's perspective. The second extension is identifying pharmaceutical technological opportunities based on multilayer networks. The last extension is integrating multi-source data in multilayer networks. Then, the proposed framework is applied to hyperuricemia drugs to prove their efficiency and effectiveness. This framework helps explain the complex relationships among diseases, ADRs, drugs, and genes. Furthermore, the analysis results provide interesting insight into pharmaceutical TOI and extend the multilayer network theory.