Department of Chemistry, KAIST, Daejeon, 34141, Korea.
Department of Biology and Chemistry, Changwon National University, Changwon, 51140, Korea.
Chem Asian J. 2022 Aug 15;17(16):e202200269. doi: 10.1002/asia.202200269. Epub 2022 Jul 20.
Most graph neural networks (GNNs) in deep-learning chemistry collect and update atom and molecule features from the fed atom (and, in some cases, bond) features, basically based on the two-dimensional (2D) graph representation of 3D molecules. However, the 2D-based models do not faithfully represent 3D molecules and their physicochemical properties, exemplified by the overlooked field effect that is a "through-space" effect, not a "through-bond" effect. We propose a GNN model, denoted as MolNet, which accommodates the 3D non-bond information in a molecule, via a noncovalent adjacency matrix , and also bond-strength information from a weighted bond matrix . Comparative studies show that MolNet outperforms various baseline GNN models and gives a state-of-the-art performance in the classification task of BACE dataset and regression task of ESOL dataset. This work suggests a future direction for the construction of deep-learning models that are chemically intuitive and compatible with the existing chemistry concepts and tools.
大多数深度学习化学中的图神经网络（GNN）从输入的原子（以及在某些情况下的键）特征中收集和更新原子和分子特征，基本上基于 3D 分子的二维（2D）图表示。然而，基于 2D 的模型不能忠实地表示 3D 分子及其物理化学性质，例如被忽视的场效应是一种“隔空”效应，而不是“键间”效应。我们提出了一个图神经网络模型，称为 MolNet，它通过非共价邻接矩阵 和来自加权键矩阵 的键强度信息来适应分子中的 3D 非键信息。比较研究表明，MolNet 优于各种基线 GNN 模型，并在 BACE 数据集的分类任务和 ESOL 数据集的回归任务中达到了最先进的性能。这项工作为构建具有化学直观性且与现有化学概念和工具兼容的深度学习模型指明了一个未来的方向。
