Abstract:Graph neural networks (GNNs) demonstrate great performance in compound property and activity prediction due to their capability to efficiently learn complex molecular graph structures. However, two main limitations persist including compound representation and model interpretability. While atom-level molecular graph representations are commonly used because of their ability to capture natural topology, they may not fully express important substructures or functional groups which significantly influence molecular properties. Consequently, recent research proposes alternative representations employing reduction techniques to integrate higher-level information and leverages both representations for model learning. However, there is still a lack of study about different molecular graph representations on model learning and interpretation. Interpretability is also crucial for drug discovery as it can offer chemical insights and inspiration for optimization. Numerous studies attempt to include model interpretation to explain the rationale behind predictions, but most of them focus solely on individual prediction with little analysis of the interpretation on different molecular graph representations. This research introduces multiple molecular graph representations that incorporate higher-level information and investigates their effects on model learning and interpretation from diverse perspectives. The results indicate that combining atom graph representation with reduced molecular graph representation can yield promising model performance. Furthermore, the interpretation results can provide significant features and potential substructures consistently aligning with background knowledge. These multiple molecular graph representations and interpretation analysis can bolster model comprehension and facilitate relevant applications in drug discovery.

Composite Graph Neural Networks for Molecular Property Prediction

Graph neural networks for the prediction of molecular structure-property relationships

Cross-dependent graph neural networks for molecular property prediction

Heterogeneous Molecular Graph Neural Networks for Predicting Molecule Properties

Multi-View Graph Neural Networks for Molecular Property Prediction

Enhancing Model Learning and Interpretation Using Multiple Molecular Graph Representations for Compound Property and Activity Prediction

Graph Neural Network Architecture Search for Molecular Property Prediction

Graph-Based Networks for Accurate Prediction of Ground and Excited State Molecular Properties from Minimal Features

Multitask Learning On Graph Neural Networks Applied To Molecular Property Predictions

Analyzing Learned Molecular Representations for Property Prediction

Bayesian Graph Neural Networks for Molecular Property Prediction

Gated Graph Recursive Neural Networks for Molecular Property Prediction

Describe Molecules by a Heterogeneous Graph Neural Network with Transformer-like Attention for Supervised Property Predictions

Graph Neural Network-Based Molecular Property Prediction with Patch Aggregation

Enhancing property and activity prediction and interpretation using multiple molecular graph representations with MMGX

GMPP-NN: a deep learning architecture for graph molecular property prediction

Graph neural networks with molecular segmentation for property prediction and structure–property relationship discovery

Extended study on atomic featurization in graph neural networks for molecular property prediction

Chain-aware graph neural networks for molecular property prediction

Comparison of Atom Representations in Graph Neural Networks for Molecular Property Prediction