ORIGINAL RESEARCH article
Front. Toxicol.
Sec. Computational Toxicology and Informatics
MoltiTox: A MultiModal Fusion Model for Molecular Toxicity Prediction
Provisionally accepted
- Sungkyunkwan University, Jongno-gu, Republic of Korea
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
We introduce MoltiTox, a novel multimodal fusion model for molecular toxicity prediction that overcomes the limitations of single-modality approaches in drug discovery. MoltiTox integrates four complementary data types: molecular graphs, SMILES strings, 2D images, and 13C NMR spectra. The model processes these inputs using four modality-specific encoders, including a Graph Isomorphism Network, a Transformer, a 2D CNN, and a 1D CNN. These heterogeneous embeddings are fused through an attention-based mechanism, allowing the model to capture complementary structural and chemical information from multiple molecular perspectives. Evaluated on the Tox21 benchmark across 12 endpoints, MoltiTox achieves a ROC-AUC of 0.831, outperforming its single-modality counterparts. These findings highlight that integrating diverse molecular representations enhances both the robustness and generalizability of toxicity prediction models. Beyond predictive performance, the inclusion of 13C NMR data offers complementary chemical insights that are not fully captured by structure-or language-based representations, suggesting its potential contribution to mechanistic understanding of molecular toxicity. By demonstrating how multimodal integration enriches molecular representations and enhances the interpretability of toxicity mechanisms, MoltiTox provides an extensible framework for developing more reliable models in computational toxicology.
Keywords: multimodal learning, deep learning, Toxicity prediction, Tox21, 13C NMR spectra, Drug Discovery, Cheminformatics
Received: 08 Oct 2025; Accepted: 21 Nov 2025.
Copyright: © 2025 Park and Lee. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Sujee Lee, sujeelee@skku.edu
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.