Chemical-Reaction-Aware Molecule Representation Learning
Chemical-Reaction-Aware Molecule Representation Learning
复制标题
DOI：
--
复制
发表时间：
2021-09
期刊：
影响因子：
--
通讯作者：
Hongwei Wang;Weijian Li;Xiaomeng Jin;Kyunghyun Cho;Heng Ji;Jiawei Han;M. Burke
中科院分区：
文献类型：
--
作者：
Hongwei Wang;Weijian Li;Xiaomeng Jin;Kyunghyun Cho;Heng Ji;Jiawei Han;M. Burke
分子表示学习（MRL）方法旨在将分子嵌入到真实的向量空间中。然而，现有的基于SMILES （Simplified molecule - input Line-Entry System）或基于GNN （Graph Neural Networks）的MRL方法要么以SMILES字符串作为输入，难以编码分子结构信息，要么过分强调GNN架构的重要性，而忽视了其泛化能力。在这里，我们建议使用化学反应来帮助学习分子表征。我们的方法的关键思想是保持分子在嵌入空间中的化学反应的等价性，即，迫使每个化学方程的反应物嵌入和产物嵌入的总和相等。事实证明，这种约束对1)保持嵌入空间的有序性和2)提高分子嵌入的泛化能力是有效的。此外，我们的模型可以使用任何GNN作为分子编码器，因此与GNN架构无关。实验结果表明，我们的方法在各种下游任务中实现了最先进的性能，例如，与最佳基线方法相比，化学反应预测的绝对Hit@1增益为17.4%，分子性质预测的绝对AUC增益为2.3%，图编辑距离预测的相对RMSE增益为18.5%。代码可在https://github.com/hwwang55/MolR上获得。
Molecule representation learning (MRL) methods aim to embed molecules into a real vector space. However, existing SMILES-based (Simplified Molecular-Input Line-Entry System) or GNN-based (Graph Neural Networks) MRL methods either take SMILES strings as input that have difficulty in encoding molecule structure information, or over-emphasize the importance of GNN architectures but neglect their generalization ability. Here we propose using chemical reactions to assist learning molecule representation. The key idea of our approach is to preserve the equivalence of molecules with respect to chemical reactions in the embedding space, i.e., forcing the sum of reactant embeddings and the sum of product embeddings to be equal for each chemical equation. This constraint is proven effective to 1) keep the embedding space well-organized and 2) improve the generalization ability of molecule embeddings. Moreover, our model can use any GNN as the molecule encoder and is thus agnostic to GNN architectures. Experimental results demonstrate that our method achieves state-of-the-art performance in a variety of downstream tasks, e.g., 17.4% absolute Hit@1 gain in chemical reaction prediction, 2.3% absolute AUC gain in molecule property prediction, and 18.5% relative RMSE gain in graph-edit-distance prediction, respectively, over the best baseline method. The code is available at https://github.com/hwwang55/MolR.