UAlign: Pushing the Limit of Template-free Retrosynthesis Prediction with Unsupervised SMILES Alignment

TL;DR

UAlign tackles template-free retrosynthesis by pairing a graph encoder (EGAT+) that leverages bond information with a Transformer decoder, guided by an unsupervised SMILES alignment that preserves shared substructures between products and reactants. The method introduces two-stage training and on-the-fly data augmentation to bridge graph and SMILES representations and to handle non-canonical SMILES. It achieves state-of-the-art performance among template-free methods and competitive results with template-based approaches across USPTO benchmarks, with notable gains in top-k accuracy and SMILES validity. The work advances template-free retrosynthesis by exploiting graph topology and alignment-driven generation, enabling more robust single-step planning and potential improvements for multi-step synthesis planning.

Abstract

Motivation: Retrosynthesis planning poses a formidable challenge in the organic chemical industry. Single-step retrosynthesis prediction, a crucial step in the planning process, has witnessed a surge in interest in recent years due to advancements in AI for science. Various deep learning-based methods have been proposed for this task in recent years, incorporating diverse levels of additional chemical knowledge dependency. Results: This paper introduces UAlign, a template-free graph-to-sequence pipeline for retrosynthesis prediction. By combining graph neural networks and Transformers, our method can more effectively leverage the inherent graph structure of molecules. Based on the fact that the majority of molecule structures remain unchanged during a chemical reaction, we propose a simple yet effective SMILES alignment technique to facilitate the reuse of unchanged structures for reactant generation. Extensive experiments show that our method substantially outperforms state-of-the-art template-free and semi-template-based approaches. Importantly, our template-free method achieves effectiveness comparable to, or even surpasses, established powerful template-based methods. Scientific contribution: We present a novel graph-to-sequence template-free retrosynthesis prediction pipeline that overcomes the limitations of Transformer-based methods in molecular representation learning and insufficient utilization of chemical information. We propose an unsupervised learning mechanism for establishing product-atom correspondence with reactant SMILES tokens, achieving even better results than supervised SMILES alignment methods. Extensive experiments demonstrate that UAlign significantly outperforms state-of-the-art template-free methods and rivals or surpasses template-based approaches, with up to 5\% (top-5) and 5.4\% (top-10) increased accuracy over the strongest baseline.

Figures (5)

• Figure 1: Overview of UAlign: Given a product molecule graph $P$ and one of its DFS order $O_P$, the graph is first fed into the graph neural network called EGAT$^+$ to obtain node features $H$. Then the positional encoding is added to $H$ according to the given DFS order $O_P$ to generate the order-aware node features $\hat{H}$. Finally the decoder takes $\hat{H}$ as input and generate the SMILES of reactants auto-regressively.
• Figure 2: An example of the process to generate order-preserving reactants SMILES. The atom mapping numbers shown on the figure are included only for clearer explanation and will be removed in our implementation to prevent any label leakage.
• Figure 3: Effects of different modules on retrosynthesis performance in reaction class unknown setting of USPTO-50K dataset. Best performance is in bold.
• Figure 4: Visualization of cross-attention over order-aware node features and the predicted tokens. The number on the y-axis is the map number of atoms in the product. The reactants atoms that not appear in product is colored red in the x-axis. $\circ$ represents the end token.
• Figure 5: Multistep retrosynthesis predictions by our method. (a) Mitapivat (b) Pacritinib citrate (c) Daprodust. The reaction centers and leaving groups are highlighted in different colors. The pathway pf molecule (a) and (b) come from literature, while the last one is verified by chemical experts.