EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

MHAFR-DDI: a multimodal hierarchical attention fusion and relation-aware architecture for drug–drug interaction event prediction

Mengli Li iD, Chao Cao iD, Quan Zou iD, Leyi Wei, Yansu Wang

Briefings in Bioinformatics · 07.01.2026 · 27 (1)

DOI: 10.1093/bib/bbag077 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Briefings in Bioinformatics, online erschienen am 23.02.2026. Verfasst von Mengli Li und 4 weiteren Autor:innen. Beteiligte Einrichtung: Centre for Artificial Intelligence Driven Drug Discovery, Faculty of Applied Science, Macao Polytechnic University, Rua de Luís Gonzaga Gomes, Macao …. Hauptthemen laut MEDLINE: Deep Learning. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 50 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0). Gefördert von: National Natural Science Foundation of China, Macao Polytechnic University.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Drug–drug interactions (DDIs) can lead to severe adverse reactions, and accurate prediction of DDI events is crucial for ensuring the safety of combination therapies and supporting drug development. Although deep learning–based approaches have achieved promising progress, existing models remain limited in modeling local–global dependencies, integrating multimodal information, and capturing cross-level molecular relationships. To address these challenges, we propose Multi-modal Hierarchical Attention Fusion and Relation-aware Architecture for DDI Event Prediction (MHAFR-DDI), a multimodal hierarchical attention fusion and relation-aware framework that enables unified modeling from intra-molecular representation to inter-molecular interaction. MHAFR-DDI adopts a two-stage pretraining–finetuning paradigm. In the pretraining stage, the model learns complementary representations from molecular sequences, 2D topological structures, and 3D spatial conformations, with modality-specific encoding mechanisms designed to capture both local structural characteristics and global semantic dependencies. Localized chemical primitives within each modality are first stabilized and then integrated into higher-level representations to ensure intra-modality stability and representational completeness. Subsequently, by introducing attention-guided data augmentation and multi-level contrastive learning, the model establishes alignment constraints across different modalities and their augmented views, thereby achieving cross-modal semantic consistency and effectively alleviating data sparsity. During the finetuning stage, the pretrained molecular representations are hierarchically fused and propagated over the drug–drug interaction graph, enabling interaction-aware information sharing among drugs and improving prediction reliability for rare drugs and long-tail interaction types. Experiments on benchmarks with 65 and 86 DDI types show that MHAFR-DDI outperforms state-of-the-art methods under the standard split, achieving macro-F1 gains of 9.5% and 6.7%, while remaining robust in weakly supervised long-tail and cold-start settings.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Deep LearningDrug InteractionsHumansAlgorithms

Stichwörter der Autor:innen

contrastive learningcross-modal representation alignmentdrug–drug interaction (DDI)multimodal fusionrelational graph learning

Autor:innen und Institutionen

  • Centre for Artificial Intelligence Driven Drug Discovery, Faculty of Applied Science, Macao Polytechnic University, Rua de Luís Gonzaga Gomes, Macao 999078, China.
  • Yangtze Delta Region Institute (Quzhou), University of Electronic Science and Technology of China, No. 1 Chengdian Road, Kecheng District, Quzhou, Zhejiang 324003, China.
  • Institute of Fundamental and Frontier Sciences, University of Electronic Science and Technology of China, No. 2006 Xiyuan Avenue, West Hi-Tech Zone, Chengdu, Sichuan 610054, China.

ORCID-Profile: Mengli Li, Chao Cao, Quan Zou

Förderung

  • National Natural Science Foundation of China
  • Macao Polytechnic University

Zitieren

APA

Li, M., Cao, C., Zou, Q., Wei, L., & Wang, Y. (2026). MHAFR-DDI: a multimodal hierarchical attention fusion and relation-aware architecture for drug–drug interaction event prediction. Briefings in Bioinformatics, 27(1). https://doi.org/10.1093/bib/bbag077

Vancouver

Li M, Cao C, Zou Q, Wei L, Wang Y. MHAFR-DDI: a multimodal hierarchical attention fusion and relation-aware architecture for drug–drug interaction event prediction. Briefings in Bioinformatics. 2026;27(1). doi:10.1093/bib/bbag077

BibTeX
@article{Li2026,
  author = {Li, Mengli and Cao, Chao and Zou, Quan and Wei, Leyi and Wang, Yansu},
  title = {MHAFR-DDI: a multimodal hierarchical attention fusion and relation-aware architecture for drug–drug interaction event prediction},
  journal = {Briefings in Bioinformatics},
  year = {2026},
  volume = {27},
  number = {1},
  doi = {10.1093/bib/bbag077}
}
Kostenlos starten →