A Statistical Mechanical Approach to Combinatorial Chemistry

Abstract

An analogy between combinatorial chemistry and Monte Carlo computer simulation is pursued. Examples of how to design libraries for both materials discovery and protein molecular evolution are given. For materials discovery, the concept of library redesign, or the use previous experiments to guide the design of new experiments, is introduced. For molecular evolution, examples of how to use ``biased'' Monte Carlo to search the protein sequence space are given. Chemical information, whether intuition, theoretical calculations, or database statistics, can be naturally incorporated as an a priori bias in the Monte Carlo approach to library design in combinatorial chemistry. In this sense, combinatorial chemistry can be viewed as an extension of traditional chemical synthesis.