Genome-wide-scale prediction of compound−protein interactions using foundation and language models based on three-dimensional structures of compounds and proteins
Bioinformatics · 2026
DOI: 10.1093/bioinformatics/btag710 ↗Kurz gefasst
Fachartikel in der Zeitschrift Bioinformatics, 2026. Verfasst von Yuga Moriyama und 7 weiteren Autor:innen. Beteiligte Einrichtung: Department of Complex Systems Science, Graduate School of Informatics, Nagoya University, Nagoya, Aichi 464-8601, Japan. Hauptthemen laut MEDLINE: Proteins, Drug Discovery, Computational Biology. Die Arbeit nennt 46 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: JSPS KAKENHI, JST BOOST.
Abstract (Originalsprache)
Motivation The identification of compound–protein interactions (CPIs) is crucial in the early stages of drug discovery. However, machine-learning (ML)-based methods based on one- and two-dimensional representations cannot capture important geometric information on the binding sites of CPIs, which limits their predictive accuracy. Results We propose DESTIN (dual structure-guided estimation of substance–target interaction networks), an ML framework for CPI prediction on a genome-wide scale from the three-dimensional (3D) structures of compounds and proteins. We constructed compound feature vectors that spatially reflect stable 3D conformations from a quantum–chemical perspective by fine-tuning a 3D structure-based foundation model. Furthermore, we constructed protein feature vectors leveraging sequence-based structure representations and protein language models, even for proteins that do not have 3D structures determined experimentally. CPI prediction was performed using an integrative ML model on 3D structure-aware feature vectors for compounds and proteins. The effectiveness of the proposed model was validated via rigorous benchmark tests on practical scenarios. DESTIN exhibited good predictive performance, even for structurally complex and pharmacologically important membrane proteins. These results show that DESTIN is promising for various pharmaceutical applications. Availability and Implementation All source code and data supporting this study are publicly available on GitHub and archived on Zenodo (https://doi.org/10.5281/zenodo.21759701).
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Substanzen
Autor:innen und Institutionen
- Department of Complex Systems Science, Graduate School of Informatics, Nagoya University, Nagoya, Aichi 464-8601, Japan.
- Integrated Glyco-Big Data Center (iGDATA), Institute for Glyco-core Research (iGCORE), Nagoya University, Nagoya, Aichi 464-8601, Japan.
- Division of Interdisciplinary Research and Development (R&D), Aichi Cancer Center Research Institute, Nagoya, Aichi 464-8681, Japan.
ORCID-Profile: Yuki Matsukiyo, Ryusuke Sawada
Förderung
- JSPS KAKENHI
- JST BOOST
Zitieren
APA
Moriyama, Y., Matsukiyo, Y., Kimura, T., Yamamoto, H., Sakajiri, Y., Shibata, T., Sawada, R., & Yamanishi, Y. (2026). Genome-wide-scale prediction of compound−protein interactions using foundation and language models based on three-dimensional structures of compounds and proteins. Bioinformatics. https://doi.org/10.1093/bioinformatics/btag710
Vancouver
Moriyama Y, Matsukiyo Y, Kimura T, Yamamoto H, Sakajiri Y, Shibata T, et al. Genome-wide-scale prediction of compound−protein interactions using foundation and language models based on three-dimensional structures of compounds and proteins. Bioinformatics. 2026. doi:10.1093/bioinformatics/btag710
BibTeX
@article{Moriyama2026,
author = {Moriyama, Yuga and Matsukiyo, Yuki and Kimura, Tsuyoshi and Yamamoto, Hiyori and Sakajiri, Yuko and Shibata, Tomokazu and Sawada, Ryusuke and Yamanishi, Yoshihiro},
title = {Genome-wide-scale prediction of compound−protein interactions using foundation and language models based on three-dimensional structures of compounds and proteins},
journal = {Bioinformatics},
year = {2026},
doi = {10.1093/bioinformatics/btag710}
}
EUVIMED