Abstract:Accurately predicting molecular properties is a challenging but essential task in drug discovery. Recently, many mono-modal deep learning methods have been successfully applied to molecular property prediction. However, mono-modal learning is inherently limited as it relies solely on a single modality of molecular representation, which restricts a comprehensive understanding of drug molecules. To overcome the limitations, we propose a multimodal fused deep learning (MMFDL) model to leverage information from different molecular representations. Specifically, we construct a triple-modal learning model by employing Transformer-Encoder, Bidirectional Gated Recurrent Unit (BiGRU), and graph convolutional network (GCN) to process three modalities of information from chemical language and molecular graph: SMILES-encoded vectors, ECFP fingerprints, and molecular graphs, respectively. We evaluate the proposed triple-modal model using five fusion approaches on six molecule datasets, including Delaney, Llinas2020, Lipophilicity, SAMPL, BACE, and pKa from DataWarrior. The results show that the MMFDL model achieves the highest Pearson coefficients, and stable distribution of Pearson coefficients in the random splitting test, outperforming mono-modal models in accuracy and reliability. Furthermore, we validate the generalization ability of our model in the prediction of binding constants for protein-ligand complex molecules, and assess the resilience capability against noise. Through analysis of feature distributions in chemical space and the assigned contribution of each modal model, we demonstrate that the MMFDL model shows the ability to acquire complementary information by using proper models and suitable fusion approaches. By leveraging diverse sources of bioinformatics information, multimodal deep learning models hold the potential for successful drug discovery.

Insights into Deep Learning Framework for Molecular Property Prediction Based on Different Tokenization Algorithms

Understanding the Limitations of Deep Models for Molecular Property Prediction: Insights and Solutions.

Accurate Physical Property Predictions via Deep Learning

Predicting Structure‐dependent Properties Directly from the Three Dimensional Molecular Images Via Convolutional Neural Networks

An accurate and interpretable deep learning model for environmental properties prediction using hybrid molecular representations

Integrating Chemical Language and Molecular Graph in Multimodal Fused Deep Learning for Drug Property Prediction

A Convolutional Neural Network Based on Self-Attention Mechanism for Molecular Property Prediction Using Molecular Hidden Fingerprints: An efficient molecular property prediction method.

DMPNN-Bert: a deep learning architecture for molecular property prediction.

Advanced deep learning methods for molecular property prediction

A Novel Molecular Representation Learning for Molecular Property Prediction with a Multiple SMILES-Based Augmentation

A deep learning framework for predicting molecular property based on multi-type features fusion

An Architecture of Deep Learning in QSPR Modeling for the Prediction of Critical Properties Using Molecular Signatures

3D-Mol: A Novel Contrastive Learning Framework for Molecular Property Prediction with 3D Information

Impact of Domain Knowledge and Multi-Modality on Intelligent Molecular Property Prediction: A Systematic Survey

Molecular Property Prediction: A Multilevel Quantum Interactions Modeling Perspective

FG-BERT: a Generalized and Self-Supervised Functional Group-Based Molecular Representation Learning Framework for Properties Prediction.

Flexible Dual-Branched Message-Passing Neural Network for a Molecular Property Prediction

Multimodal fused deep learning for drug property prediction: Integrating chemical language and molecular graph

Insight into the Structure–Odor Relationship of Molecules: A Computational Study Based on Deep Learning

SMG-BERT: integrating stereoscopic information and chemical representation for molecular property prediction