3DGraphX: Explaining 3D Molecular Graph Models via Incorporating Chemical Priors
Conference
Liu, X, Luo, D, Gao, W et al. (2025). 3DGraphX: Explaining 3D Molecular Graph Models via Incorporating Chemical Priors
. 1 859-870. 10.1145/3690624.3709302
Liu, X, Luo, D, Gao, W et al. (2025). 3DGraphX: Explaining 3D Molecular Graph Models via Incorporating Chemical Priors
. 1 859-870. 10.1145/3690624.3709302
We consider the explanation of 3D graph neural networks (GNNs) in the field of molecular learning. Recent studies have modeled molecules as 3D graphs, but there exist formidable challenges for 3D graph explanation. In this work, we propose a novel and principled paradigm, known as 3DGraphX, for 3D molecular graph explanation. Unlike existing 2D GNN explanation methods, 3DGraphX focuses on 3D motifs, which are subgraphs showing great occurrence and function significance in molecular activities. Once generated, 3D motifs are fixed in the explanation model; hence, 3DGraphX produces more accurate and chemically plausible explanations in an efficient manner. 3DGraphX contains two branches with several novel methods for instance-level and geometry-level explanations, respectively. Two novel components, known as the mask pooling component and mask unpooling component, are developed to discover important motifs for each 3D molecule as the instance-level explanation. Local spherical coordinate systems are built to investigate the relative positions among motifs for geometry-level explanation. Altogether, 3DGraphX sheds light on the characteristics of molecules as well as the behaviors of 3D GNNs in molecular learning. Experimental results show that 3DGraphX significantly outperforms baselines in instance-level explanation with various explanation budgets. Additional experiments show that 3DGraphX reveals the important geometries taken by 3D GNNs for accurate molecular learning. The code is publicly available at https://github.com/xufliu/3DGraphX.
