A deterministic annealing approach to combinatorial library design for drug discovery

In this paper we propose a new algorithm for the design of a lead generation library for the purpose of drug discovery. The algorithm is based on a modification of the deterministic annealing algorithm used for clustering and locational optimization problems. The algorithm addresses two key criteria namely, diversity and representativeness of compounds in the obtained library. Thus the time-intensive process of finding a lead compound for drug discovery can be accelerated using this technique of library design.