Learning Chemical Reaction Representation with Reactant-Product Alignment

TL;DR

RAlign introduces a novel chemical reaction representation learning framework that explicitly models reactant–product atom alignment and reaction centers, coupled with an adapter mechanism to incorporate diverse reaction conditions. The Atom Aligned Encoder and Reaction-Center-Aware Decoder enable the model to focus on key transformation sites while integrating conditional information, yielding strong improvements across reaction-condition prediction/generation, yield, and selectivity tasks. Ablation studies and case analyses corroborate the importance of atom-wise alignment and RC-aware attention in producing robust, interpretable reaction representations. The approach holds promise for improved synthesis planning and mechanism understanding, with potential for large-scale pretraining and open-source adoption.

Abstract

Organic synthesis stands as a cornerstone of the chemical industry. The development of robust machine learning models to support tasks associated with organic reactions is of significant interest. However, current methods rely on hand-crafted features or direct adaptations of model architectures from other domains, which lack feasibility as data scales increase or ignore the rich chemical information inherent in reactions. To address these issues, this paper introduces RAlign, a novel chemical reaction representation learning model for various organic reaction-related tasks. By integrating atomic correspondence between reactants and products, our model discerns the molecular transformations that occur during the reaction, thereby enhancing comprehension of the reaction mechanism. We have designed an adapter structure to incorporate reaction conditions into the chemical reaction representation, allowing the model to handle various reaction conditions and to adapt to various datasets and downstream tasks. Additionally, we introduce a reaction-center-aware attention mechanism that enables the model to concentrate on key functional groups, thereby generating potent representations for chemical reactions. Our model has been evaluated on a range of downstream tasks. Experimental results indicate that our model markedly outperforms existing chemical reaction representation learning architectures on most of the datasets. We plan to open-source the code contingent upon the acceptance of the paper.

Figures (9)

• Figure 1: Overview of RAlign. For a given chemical reaction, the molecule graphs of reactants and products, along with their atomic correspondence, are input into a $L$-layer Atom Aligned Encoder to extract reaction node features $H^{P(L)}$ and $H^{R(L)}$. If the reaction conditions are also provided, they are encoded by a condition encoder and merged into $H^{P(L)}$ and $H^{R(L)}$ via an adapter structure proposed in this study. The resulting features are then processed by the RC-aware decoder to produce outputs for subsequent tasks. We have tailored two decoders for both sequential output and single output, both featuring an RC-aware cross-attention layer to concentrate on key reaction motifs.
• Figure 2: Different visualizations of node embeddings at each encoder layer, comparing the aligned encoder with the non-aligned encoder.
• Figure 3: Visualization of RC-aware cross attention weights of each node in reactants and products for catalyst prediction
• Figure 4: The visualization the attention weights of each node in reactants and products while predicting the reaction condition combination for the reduction of nitro groups.
• Figure 5: The visualization of the RC-aware cross attention coefficients for the reaction condition combination prediction of a amide coupling reaction.