Multiobjective optimization of biological and physical properties in drug discovery

Abstract

Drug discovery is a multiobjective optimization process whereby a range of physical, biological, safety, economic and other practical considerations must be combined together in order to identify a suitable molecule to treat a given disease. The precise balance of these parameters is not always clear due to the complexity of the science and this is reflected in the attrition rate of >90% of clinical candidates in later development. Great advances have undoubtedly been made in pharmaceutical science, however our holistic understanding of the precise combination of properties are needed to make a given drug is still limited. This has led to the deployment of sophisticated methods, processes and assays to probe the pharmacological, pharmacokinetic & dynamic as well as safety aspects. In this chapter we consider the importance of multiobjective optimization with a particular focus on physicochemical parameters, more concern earlier in lead generation/optimization, and PK/PD considerations, generally more extensively evaluated in late lead optimization. We also discuss how as an industry we can sometimes become overly focused on individual properties or metrics that are within our understanding and control, even though they may be of finite relevance. We discuss to what extent we place too much focus on some properties we understand or can predict, to the detriment of other factors, potentially stifling scientific innovation and opportunities for serendipitous discoveries.