MolGraph-xLSTM: A graph-based dual-level xLSTM framework with multi-head mixture-of-experts for enhanced molecular representation and interpretability

2025-01-30

AI Summary

Existing GNNs struggle to capture long-range dependencies in molecular graphs, limiting their performance in molecular property prediction for drug discovery. To address this, we propose a dual-scale graph neural modeling framework: at the atom level, we employ an xLSTM-enhanced GNN with skip connections to jointly model local and long-range structural patterns; at the motif level, we construct complementary graph views and fuse multi-granularity representations via a multi-head mixture-of-experts (MHMoE) module. This work introduces the first dual-level xLSTM architecture for molecular graphs, pioneers the integration of skip knowledge with xLSTM for atomic representation learning, and innovatively incorporates motif-level graph construction and MHMoE to enhance both interpretability and expressive power. Evaluated on ten molecular property prediction benchmarks, our method consistently outperforms state-of-the-art baselines: it achieves an average AUROC gain of 3.18% on classification tasks and an average RMSE reduction of 3.83% on regression tasks, with notable improvements of 7.03% on BBBP and 7.54% on ESOL.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Knowledge Representation and Reasoning: Description Logics

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Predicting molecular properties is essential for drug discovery, and computational methods can greatly enhance this process. Molecular graphs have become a focus for representation learning, with Graph Neural Networks (GNNs) widely used. However, GNNs often struggle with capturing long-range dependencies. To address this, we propose MolGraph-xLSTM, a novel graph-based xLSTM model that enhances feature extraction and effectively models molecule long-range interactions. Our approach processes molecular graphs at two scales: atom-level and motif-level. For atom-level graphs, a GNN-based xLSTM framework with jumping knowledge extracts local features and aggregates multilayer information to capture both local and global patterns effectively. Motif-level graphs provide complementary structural information for a broader molecular view. Embeddings from both scales are refined via a multi-head mixture of experts (MHMoE), further enhancing expressiveness and performance. We validate MolGraph-xLSTM on 10 molecular property prediction datasets, covering both classification and regression tasks. Our model demonstrates consistent performance across all datasets, with improvements of up to 7.03% on the BBBP dataset for classification and 7.54% on the ESOL dataset for regression compared to baselines. On average, MolGraph-xLSTM achieves an AUROC improvement of 3.18% for classification tasks and an RMSE reduction of 3.83% across regression datasets compared to the baseline methods. These results confirm the effectiveness of our model, offering a promising solution for molecular representation learning for drug discovery.

Problem

• Graph Neural Networks
• Molecular Graphs
• Drug Discovery

Innovation

• MolGraph-xLSTM
• Dual-Perspective Analysis
• Multi-Head Mixed Experts (MHMoE)