Knowledge-aware contrastive heterogeneous molecular graph learning

TL;DR

This paper introduces KCHML, a novel framework that encodes molecular graphs as heterogeneous structures with external knowledge, improving molecular property prediction and drug interaction tasks through contrastive learning.

Contribution

It presents a new heterogeneous graph encoding method with a dual message-passing mechanism and external knowledge integration, advancing molecular representation learning.

Findings

KCHML outperforms existing models in property prediction tasks.

The framework effectively captures complex molecular features.

Demonstrates superior performance in drug-drug interaction prediction.

Abstract

Molecular representation learning is pivotal in predicting molecular properties and advancing drug design. Traditional methodologies, which predominantly rely on homogeneous graph encoding, are limited by their inability to integrate external knowledge and represent molecular structures across different levels of granularity. To address these limitations, we propose a paradigm shift by encoding molecular graphs into heterogeneous structures, introducing a novel framework: Knowledge-aware Contrastive Heterogeneous Molecular Graph Learning (KCHML). This approach leverages contrastive learning to enrich molecular representations with embedded external knowledge. KCHML conceptualizes molecules through three distinct graph views-molecular, elemental, and pharmacological-enhanced by heterogeneous molecular graphs and a dual message-passing mechanism. This design offers a comprehensive representation for property prediction, as well as for downstream tasks such as drug-drug interaction (DDI) prediction. Extensive benchmarking demonstrates KCHML's superiority over state-of-the-art molecular property prediction models, underscoring its ability to capture intricate molecular features.