Multi-level Self-supervised Pretraining on Compositional Hierarchical Graph for Molecular Property Prediction

TL;DR

This paper introduces MolCHG, a multi-level self-supervised pretraining framework for molecular graphs that captures hierarchical structural information to improve molecular property prediction.

Contribution

The work proposes a novel Compositional Hierarchical Graph and multi-level pretraining objectives to better utilize bond and atom information in molecular graphs.

Findings

MolCHG outperforms baselines on seven MoleculeNet datasets.

Multi-level supervision signals are complementary and enhance performance.

Bond-level information is effectively integrated as an independent semantic layer.

Abstract

Self-supervised pretraining on molecular graphs has emerged as a promising approach for molecular property prediction, yet most existing methods operate at a single structural granularity and treat bond information as auxiliary edge attributes rather than as an independent semantic layer. In this work, we propose MolCHG, a multi-level self-supervised pretraining framework built upon a novel Compositional Hierarchical Graph that organizes molecular structure into four types of nodes across three semantic levels. By introducing a bond graph that operates in parallel with the atom graph, our architecture elevates bond-level information to independently evolving node representations, enabling fragment nodes to aggregate atom-level and bond-level semantics on an equal footing. We design three level-specific pretraining objectives: an atom-bond cross-view contrastive task that aligns the atom-view and bond-view representations within each fragment, a fragment-level functional group prediction task to inject domain-relevant chemical knowledge, and graph-level structure prediction tasks to encode global molecular topology. Experiments on nine MoleculeNet benchmarks demonstrate that MolCHG achieves the best performance on seven datasets across both classification and regression tasks, remaining competitive with the strongest baselines on the rest. Ablation studies further confirm that the multi-level supervision signals are complementary and that each component contributes to the overall performance.