Medicinal chemist structure function correlations

The presence of chemically reactive structural features in potential drug candidates, especially when caused by metabolism, has been linked to idiosyncratic toxicity [56,57] although in most cases this is hard to prove unambiguously, and there is no evidence that idiosyncratic toxicity is correlated with specific physical properties per se. The best strategy for the medicinal chemist is avoidance of the liabilities associated with inherently chemically reactive or metabolically activated functional groups [58]. For reactive metabolites, protein covalent-binding screens [59] and genetic toxicity tests (Ames) of putative metabolites, for example, embedded anilines, can be employed in risky chemical series. 

The term CADD has been used to describe two aspects of the recent use of computational tools that aid computational and medicinal chemists in the search for new drug candidates. In the first approach, medicinal chemists attempt to describe the predominant statistical correlation of biological activity with directly measurable physicochemical parameters or characteristics of drugs and is known as Quantitative Structure-Activity Relationships (QSAR). The central idea is that compounds exhibit biological activity based on structural characteristics. It should then be possible to correlate the associated biological activity with various critical parameters. Ingeneral, the biological activity may be considered a function of hydropho-bicity, electrostatics, and steric forces [Eq. (18)]. 

---