Molecular geometry prediction of flexible molecules, or conformer search, is a longstanding challenge in computational chemistry.

Molecular geometry prediction of flexible molecules, or conformer search, is a longstanding challenge in computational chemistry.

CEST researchers developed a new machine learning approach based on a low-energy latent space (LOLS) and density functional theory (DFT) to search for molecular conformers. Molecular conformer search is a topic of great importance in computational chemistry, drug design and material science. The challenge is to identify low-energy conformers in the first place. This difficulty arises from the high complexity of search spaces, as well as the computational cost associated with accurate quantum chemical methods. In the past, conformer search would take up considerable time and computational resources.

Molecular geometry prediction of flexible molecules, or conformer search, is a long-standing challenge in computational chemistry. This task is of great importance for predicting structure-activity relationships for a wide variety of substances ranging from biomolecules to ubiquitous materials. Substantial computational resources are invested in Monte Carlo and Molecular Dynamics methods to generate diverse and representative conformer sets for medium to large molecules, which are yet intractable to chemoinformatic conformer search methods. We present TorsionNet, an efficient sequential conformer search technique based on reinforcement learning under the rigid rotor approximation. The model is trained via curriculum learning, whose theoretical benefit is explored in detail, to maximize a novel metric grounded in thermodynamics called the Gibbs Score. Our experimental results show that TorsionNet outperforms the highest scoring chemoinformatics method by 4x on large branched alkanes, and by several orders of magnitude on the previously unexplored biopolymer lignin, with applications in renewable energy.