Enhanced Thompson sampling by roulette wheel selection for screening ultralarge combinatorial libraries
Hongtao Zhao, Eva Nittinger, Melissa A. Yu, Symon M. Gathiaka, W. Patrick Walters, Christian Tyrchan · Journal of Cheminformatics · 2025
Chemical space exploration has gained significant interest with the increasing availability of building blocks, enabling the creation of ultralarge virtual libraries containing billions or trillions of compounds. However, challenges remain in selecting the most suitable compounds for synthesis, especially in hit expansion. Thompson sampling, a probabilistic search method, has recently been proposed to improve efficiency by operating in reagent space rather than product space. Here, we address some of its limitations by introducing a roulette wheel selection method combined with a thermal cycling approach to balance greedy search and diversity-driven exploration. The effectiveness of this method is demonstrated through 109 queries against twenty distinct 1-million-compound libraries using ROCS. To the best of our knowledge, this is the first report of integrating variational autoencoders with normalizing flows in a comprehensive molecular design workflow. It is also the first application of conditional normalizing flows to molecular design using string based molecular representation. Our method yielded novel molecules with performance metrics surpassing those in the 2.4 million-compound ChEMBL database, highlighting its potential for identifying promising drug candidates.