Statistical analysis of the effects of common chemical substituents on ligand potency.

ABSTRACT

The results of a statistical analysis of more than 84,000 compounds from lead optimization programs against 30 different protein targets is presented, with a focus on the effects that different chemical substituents have on compound potency. It is observed that the potency changes induced by most chemical groups follows a nearly normal distribution centered near zero (i.e., no effect on potency). However, the widths of the distributions vary significantly between different substituents, and these effects cannot be rationalized by simple physicochemical parameters. In addition, certain substituents consistently bias the distribution toward higher or lower potency, suggesting the existence of preferred and nonpreferred chemical groups for lead optimization. The implications of these results for understanding protein-ligand recognition and for enhancing the efficiency and speed of lead optimization will be discussed.