(9a) Retrosynthesis Based on Chemical Similarity: Applications in Enzymatic Synthesis

Enzymatic syntheses often result in extraordinary efficiencies due to their stereo-, regio-, and enantio- selective transformations. Here, we present a computational enzymatic synthesis planner to facilitate such approaches. To address this challenge, we (1) demonstrated that molecular similarity is an effective metric to propose retrosynthetic disconnections based on analogy to precedent enzymatic reactions in UniProt/ RHEA (2) trained a neural network capable of understanding the substrate promiscuity of enzymes and evaluating the likelihood of experimental success. We successfully planned enzymatic synthesis routes for both active pharmaceutical ingredients (e.g. Islatravir) and commodity chemicals (e.g. 1,4-butanediol, 1,3-propanediol, branched-chain higher alcohols/biofuels), in a retrospective fashion. Our approach provides an important first step towards solving the challenging problem of incorporating both enzymatic- and organic chemistry- based transformations into a computer aided synthesis planning workflow.