EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Hypergraph learning with multi-dimensional metabolite feature extractions and static–dynamic attention mechanisms to fill missing reactions in metabolic networks

Kai Wang, Jiajun Qu, Fei Liu, Xiaoli Luan, Jingwen Zhou

Briefings in Bioinformatics · 04.05.2026 · 27 (3)

DOI: 10.1093/bib/bbag314 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Briefings in Bioinformatics, online erschienen am 17.06.2026. Verfasst von Kai Wang und 4 weiteren Autor:innen. Beteiligte Einrichtung: Key Laboratory of Advanced Process Control for Light Industry (Ministry of Education), School of Automation and Intelligent Science (School of Intern…. Hauptthemen laut MEDLINE: Metabolic Networks and Pathways, Machine Learning, Models, Biological, Computational Biology. Die Arbeit nennt 48 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Key Research and Development Program of China, National Natural Science Foundation of China, National First-class Discipline Program of Light Industry Technology and Engineering.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Genome-scale metabolic models (GEMs) can effectively facilitate many fields in synthetic biology, biomanufacturing, and biomedicine. Reconstructing high-quality GEMs is crucial for accurate phenotype predictions of organisms. However, draft GEMs generated by automated reconstruction tools contain many knowledge gaps, especially missing reactions. The existing machine learning-based gap-filling approaches need to be further developed. In this article, we propose a novel HyperGraph Learning approach with Multi-dimensional metabolite feature extractions and static–dynamic Attention mechanisms (HGLMA) for predicting and teasing out missing reactions in GEM gap-fillings. HGLMA simultaneously uses two pretrained language models to proceed multi-dimensional metabolite feature extractions, which are further fused and regarded as node embeddings for graph learning. The directed and high-order associations between metabolites in reactions of GEMs are deeply mined by successively employing a directional graph network and a hypergraph neural network. Before outputting the predicted confidence score for candidate reactions, the static–dynamic multi-head attention mechanism is utilized to automatically learn attention weights and to identify key metabolites within any candidate reaction. The five-fold cross-validation results on 108 BiGG GEMs show that HGLMA significantly outperforms other state-of-the-art machine learning-based approaches both in prediction performances and in the ability of discovering missing reactions from metabolic reaction pools. The ablation study shows the contributions of multi-dimensional feature extractions and static–dynamic attention mechanisms. In addition, the phenotype prediction results of 24 bacterial organisms demonstrate the effectiveness and superiority of gap-fillings by HGLMA.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Metabolic Networks and PathwaysMachine LearningModels, BiologicalComputational BiologyGraph Neural NetworksAlgorithms

Stichwörter der Autor:innen

gap-fillinggenome-scale metabolic modelhypergraph learningpretrained language modelstatic–dynamic attention mechanism

Autor:innen und Institutionen

  • Key Laboratory of Advanced Process Control for Light Industry (Ministry of Education), School of Automation and Intelligent Science (School of Internet of Things), Jiangnan University, 1800 Lihu Road, Wuxi, Jiangsu 2141…
  • Science Center for Future Foods, Jiangnan University, 1800 Lihu Road, Wuxi, Jiangsu 214122, China.
  • Key Laboratory of Industrial Biotechnology, Ministry of Education and School of Biotechnology, Jiangnan University, 1800 Lihu Road, Wuxi, Jiangsu 214122, China.
  • Engineering Research Center of Ministry of Education on Food Synthetic Biotechnology, Jiangnan University, 1800 Lihu Road, Wuxi, Jiangsu 214122, China.
  • Jiangsu Province Engineering Research Center of Food Synthetic Biotechnology, Jiangnan University, 1800 Lihu Road, Wuxi, Jiangsu 214122, China.

Förderung

  • National Key Research and Development Program of China
  • National Natural Science Foundation of China
  • National First-class Discipline Program of Light Industry Technology and Engineering

Zitieren

APA

Wang, K., Qu, J., Liu, F., Luan, X., & Zhou, J. (2026). Hypergraph learning with multi-dimensional metabolite feature extractions and static–dynamic attention mechanisms to fill missing reactions in metabolic networks. Briefings in Bioinformatics, 27(3). https://doi.org/10.1093/bib/bbag314

Vancouver

Wang K, Qu J, Liu F, Luan X, Zhou J. Hypergraph learning with multi-dimensional metabolite feature extractions and static–dynamic attention mechanisms to fill missing reactions in metabolic networks. Briefings in Bioinformatics. 2026;27(3). doi:10.1093/bib/bbag314

BibTeX
@article{Wang2026,
  author = {Wang, Kai and Qu, Jiajun and Liu, Fei and Luan, Xiaoli and Zhou, Jingwen},
  title = {Hypergraph learning with multi-dimensional metabolite feature extractions and static–dynamic attention mechanisms to fill missing reactions in metabolic networks},
  journal = {Briefings in Bioinformatics},
  year = {2026},
  volume = {27},
  number = {3},
  doi = {10.1093/bib/bbag314}
}
Kostenlos starten →