Abstract

Combinatorial library design is by nature a multicriterion problem. These criteria often include reagent diversity, product similarity to lead compounds and product novelty with respect to a corporate compound bank. More recently, developability and druglikeness have also attracted much attention in library design practices. To address this multicriterion design problem, we have developed a computer program (PICCOLO) that simultaneously optimizes all the factors under consideration using a weighted sum optimization technique. In this paper, we describe the overall design of this program and the formulation of individual penalty functions that characterize the underlying design criteria. We also give an example to illustrate the process and the result of a library design using this program.