EnGCI: enhancing GPCR-compound interaction prediction via large molecular models and KAN network.

Publisher: BioMed Central Country of Publication: England NLM ID: 101190720 Publication Model: Electronic Cited Medium: Internet ISSN: 1741-7007 (Electronic) Linking ISSN: 17417007 NLM ISO Abbreviation: BMC Biol Subsets: MEDLINE

Original Publication: [London] : BioMed Central, c2003-

Receptors, G-Protein-Coupled*/chemistry ; Receptors, G-Protein-Coupled*/metabolism ; Drug Discovery*/methods ; Neural Networks, Computer* ; Models, Molecular*; Deep Learning ; Machine Learning

Competing Interests: Declarations. Ethics approval and consent to participate: Not applicable. Consent for publication: Not applicable. Competing interests: The authors declare no competing interests.
Background: Identifying GPCR-compound interactions (GCI) plays a significant role in drug discovery and chemogenomics. Machine learning, particularly deep learning, has become increasingly influential in this domain. Large molecular models, due to their ability to capture detailed structural and functional information, have shown promise in enhancing the predictive accuracy of downstream tasks. Consequently, exploring the performance of these models in GCI prediction, as well as evaluating their effectiveness when integrated with other deep learning models, has emerged as a compelling research area. This paper aims to investigate these challenges.
Results: This study introduces EnGCI, a novel model comprising two distinct modules. The MSBM integrates a graph isomorphism network (GIN) and a convolutional neural network (CNN) to extract features from GPCRs and compounds, respectively. These features are then processed by a Kolmogorov-Arnold network (KAN) for decision-making. The LMMBM utilizes two large-scale pre-trained models to extract features from compounds and GPCRs, and subsequently, KAN is again employed for decision-making. Each module leverages different sources of multimodal information, and their fusion enhances the overall accuracy of GPCR-compound interaction (GCI) prediction. Evaluating the EnGCI model on a rigorously curated GCI dataset, we achieved an AUC of approximately 0.89, significantly outperforming current state-of-the-art benchmark models.
Conclusions: The EnGCI model integrates two complementary modules: one that learns molecular features from scratch for the GPCR-compound interaction (GCI) prediction task, and another that extracts molecular features using pre-trained large molecular models. After further processing and integration, these multimodal information sources enable a more profound exploration and understanding of the complex interaction relationships between GPCRs and compounds. The EnGCI model offers a robust and efficient framework that enhances GCI predictive capabilities and has the potential to significantly contribute to GPCR drug discovery.
(© 2025. The Author(s).)

Nature. 2017 May 18;545(7654):317-322. (PMID: 28489817)
Nature. 2011 Jul 19;477(7366):549-55. (PMID: 21772288)
Front Bioinform. 2023 May 26;3:1193025. (PMID: 37304403)
BMC Med Inform Decis Mak. 2020 Mar 18;20(Suppl 2):49. (PMID: 32183788)
Curr Opin Struct Biol. 2008 Dec;18(6):734-40. (PMID: 18957321)
Proteins. 2016 Sep;84(9):1224-33. (PMID: 27191053)
PLoS Comput Biol. 2024 Jul 22;20(7):e1012248. (PMID: 39038042)
Crit Rev Clin Lab Sci. 2005;42(4):311-92. (PMID: 16281738)
Bioinformatics. 2020 Aug 15;36(16):4406-4414. (PMID: 32428219)
Mol Inform. 2020 May;39(5):e1900062. (PMID: 32003548)
J Chem Inf Model. 2013 Jun 24;53(6):1253-62. (PMID: 23721295)
Comput Struct Biotechnol J. 2022 May 18;20:2564-2573. (PMID: 35685352)
Brief Bioinform. 2022 Nov 19;23(6):. (PMID: 36242566)
Biochem Pharmacol. 2015 Dec 15;98(4):541-55. (PMID: 26265138)
Comput Biol Chem. 2016 Feb;60:59-71. (PMID: 26674225)
Bioinformatics. 2021 May 23;37(8):1140-1147. (PMID: 33119053)
J Proteome Res. 2017 Apr 7;16(4):1401-1409. (PMID: 28264154)
PLoS Comput Biol. 2019 Jun 14;15(6):e1007129. (PMID: 31199797)
Brief Bioinform. 2023 Sep 22;24(6):. (PMID: 37930025)
Brief Bioinform. 2024 May 23;25(4):. (PMID: 38864340)
Nature. 2013 Feb 14;494(7436):185-94. (PMID: 23407534)
BMC Bioinformatics. 2020 Apr 20;21(1):150. (PMID: 32312232)
ACS Omega. 2021 Jan 21;6(4):3037-3045. (PMID: 33553921)
BMC Bioinformatics. 2022 Aug 18;23(Suppl 9):346. (PMID: 35982407)
Front Cell Dev Biol. 2021 Feb 01;8:623858. (PMID: 33598456)
Nat Commun. 2024 May 28;15(1):4536. (PMID: 38806453)
Cell. 2016 Aug 11;166(4):907-919. (PMID: 27499021)
Brief Bioinform. 2024 Jan 22;25(2):. (PMID: 38279650)
J Integr Bioinform. 2020 Dec 29;18(2):155-165. (PMID: 34171942)
PLoS One. 2013 Aug 27;8(8):e72234. (PMID: 24015221)
BMC Bioinformatics. 2022 Jan 20;22(Suppl 12):461. (PMID: 35057737)
Biochemistry. 2017 Oct 24;56(42):5628-5634. (PMID: 28967738)
Nature. 2009 May 21;459(7245):356-63. (PMID: 19458711)
Methods. 2024 Mar;223:75-82. (PMID: 38286333)
Nat Commun. 2021 Sep 15;12(1):5465. (PMID: 34526500)
Nat Rev Drug Discov. 2017 Dec;16(12):829-842. (PMID: 29075003)
Nat Rev Drug Discov. 2008 Apr;7(4):339-57. (PMID: 18382464)
PLoS One. 2024 Aug 22;19(8):e0309078. (PMID: 39172871)
Proc Natl Acad Sci U S A. 2019 Feb 19;116(8):3288-3293. (PMID: 30728297)

2024AFD038 Natural Science Foundation of Hubei Province in China; 2023AFB042 Natural Science Foundation of Hubei Province in China; 2024AFD038 Natural Science Foundation of Hubei Province in China; 62306108 National Natural Science Foundation of China; G2023027007L National Science and Technology Ministry Projects

Keywords: GPCR-compound interaction prediction; Graph isomorphism network; Kolmogorov-Arnold network; Large molecular models

0 (Receptors, G-Protein-Coupled)

Date Created: 20250515 Date Completed: 20250516 Latest Revision: 20250518

20250519

PMC12082873

10.1186/s12915-025-02238-3

40375308