RetroDiff: Retrosynthesis as Multi-stage Distribution Interpolation

TL;DR

RetroDiff introduces a multi-stage diffusion approach for retrosynthesis, effectively modeling the process as a graph-to-graph generative task that outperforms existing methods in accuracy and validity.

Contribution

The paper presents RetroDiff, a novel diffusion-based framework that decomposes retrosynthesis into multi-stage graph generation, integrating chemical reaction templates with diffusion models.

Findings

RetroDiff surpasses semi-template methods in accuracy.

RetroDiff outperforms template-based methods in large-scale scenarios.

RetroDiff achieves higher molecular validity compared to template-free methods.

Abstract

Retrosynthesis poses a key challenge in biopharmaceuticals, aiding chemists in finding appropriate reactant molecules for given product molecules. With reactants and products represented as 2D graphs, retrosynthesis constitutes a conditional graph-to-graph (G2G) generative task. Inspired by advancements in discrete diffusion models for graph generation, we aim to design a diffusion-based method to address this problem. However, integrating a diffusion-based G2G framework while retaining essential chemical reaction template information presents a notable challenge. Our key innovation involves a multi-stage diffusion process. We decompose the retrosynthesis procedure to first sample external groups from the dummy distribution given products, then generate external bonds to connect products and generated groups. Interestingly, this generation process mirrors the reverse of the widely adapted semi-template retrosynthesis workflow, \emph{i.e.} from reaction center identification to synthon completion. Based on these designs, we introduce Retrosynthesis Diffusion (RetroDiff), a novel diffusion-based method for the retrosynthesis task. Experimental results demonstrate that RetroDiff surpasses all semi-template methods in accuracy, and outperforms template-based and template-free methods in large-scale scenarios and molecular validity, respectively. Code: https://github.com/Alsace08/RetroDiff.