EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

DrugBLIP: exploring the protein–molecule interaction mechanisms with a multi-task learning graph transformer

Rubo Wang, Xingyu Gao iD, Peilin Zhao

Bioinformatics · 07.04.2026 · 42 (4)

DOI: 10.1093/bioinformatics/btag069 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Bioinformatics, online erschienen am 10.04.2026. Verfasst von Rubo Wang und 2 weiteren Autor:innen. Beteiligte Einrichtung: Institute of Microelectronics, Chinese Academy of Sciences, Beijing, 100029, China. Hauptthemen laut MEDLINE: Proteins, Molecular Docking Simulation, Deep Learning, Software. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 27 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Science and Technology Innovation (STI) 2030-Major Projects, National Natural Science Foundation of China, State Key Laboratory of Chemo and Biosensing (Hunan University).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Motivation Traditional drug discovery methods are costly and inefficient, while existing deep learning approaches remain limited by task specificity and practical applicability. Accurately modeling protein–molecule interactions is critical for advancing virtual screening, docking, and drug design. Results We propose DrugBLIP, a multi-task graph transformer model based on SE(3)-equivariant architectures, to unify protein–molecule interaction learning. By integrating contrastive learning, matching tasks, and docking optimization, DrugBLIP captures 3D spatial relationships through a hybrid graph transformer framework. Evaluations demonstrate state-of-the-art performance: DrugBLIP achieves an AUROC of 0.8217 and BEDROC of 0.5743 on virtual screening, outperforming traditional and deep learning baselines by 10%–127% across metrics. It also attains 91.2% top-1 docking success on CASF-2016 and 41.8% target fishing accuracy, showcasing robustness in diverse scenarios. Additionally, DrugBLIP reduces computational time by 700× compared to traditional docking tools. Availability and implementation Code is available at https://github.com/Wolkenwandler/DrugBLIP and archived at Zenodo with DOI: 10.5281/zenodo.16990700.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

ProteinsMolecular Docking SimulationDeep LearningSoftwareDrug DiscoveryComputational BiologyProtein Binding

Substanzen

Proteins

Autor:innen und Institutionen

  • Institute of Microelectronics, Chinese Academy of Sciences, Beijing, 100029, China.
  • University of Chinese Academy of Sciences, Beijing, 100049, China.
  • School of Artificial Intelligence, Shanghai Jiao Tong University, Shanghai, 200230, China.

ORCID-Profile: Xingyu Gao

Förderung

  • Science and Technology Innovation (STI) 2030-Major Projects
  • National Natural Science Foundation of China
  • State Key Laboratory of Chemo and Biosensing (Hunan University)
  • Science and Technology Innovation (STI) 2030—Major Projects

Zitieren

APA

Wang, R., Gao, X., & Zhao, P. (2026). DrugBLIP: exploring the protein–molecule interaction mechanisms with a multi-task learning graph transformer. Bioinformatics, 42(4). https://doi.org/10.1093/bioinformatics/btag069

Vancouver

Wang R, Gao X, Zhao P. DrugBLIP: exploring the protein–molecule interaction mechanisms with a multi-task learning graph transformer. Bioinformatics. 2026;42(4). doi:10.1093/bioinformatics/btag069

BibTeX
@article{Wang2026,
  author = {Wang, Rubo and Gao, Xingyu and Zhao, Peilin},
  title = {DrugBLIP: exploring the protein–molecule interaction mechanisms with a multi-task learning graph transformer},
  journal = {Bioinformatics},
  year = {2026},
  volume = {42},
  number = {4},
  doi = {10.1093/bioinformatics/btag069}
}
Kostenlos starten →