{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,11,4]],"date-time":"2025-11-04T18:36:51Z","timestamp":1762281411574,"version":"build-2065373602"},"reference-count":50,"publisher":"Public Library of Science (PLoS)","issue":"11","license":[{"start":{"date-parts":[[2025,11,4]],"date-time":"2025-11-04T00:00:00Z","timestamp":1762214400000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100000038","name":"Natural Sciences and Engineering Research Council of Canada","doi-asserted-by":"publisher","award":["RGPIN-2023-05224"],"award-info":[{"award-number":["RGPIN-2023-05224"]}],"id":[{"id":"10.13039\/501100000038","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.plosone.org"],"crossmark-restriction":false},"short-container-title":["PLoS One"],"abstract":"<jats:p>The analysis of molecular interactions between antigens and antibodies is crucial for understanding the immunological mechanisms underlying the immune response and for developing effective therapies against various diseases. In this context, the ability to distinguish between protein interfaces that form stable and unstable complexes is a key step in the design of therapeutic antibodies and vaccines. In recent years, deep learning models have provided advanced tools for biomedical research. This work introduces a novel approach to analyzing antibody-antigen interactions, and in particular SARS-CoV-2 spike protein-targeting antibodies, using a Siamese Neural Network specifically designed to integrate depth maps with geometric descriptors of molecular surfaces. By combining these representations, the model captures geometrical shape complementarity to differentiate between stable and unstable protein complexes. The network was trained using image-based representations of antigens and antibodies interfaces enriched with geometric descriptors, using data that include binders and non-binders of the SARS-CoV-2 spike protein receptor-binding domain. The deep learning network operates by comparing feature vectors representing these molecular surfaces; pairs with closer vectors in feature space are associated with stable interactions, while those with more distant vectors suggest instability. Extensive testing with different configurations achieved an accuracy of 90%, demonstrating the robustness of this approach to predict interactions. This innovative integration of artificial intelligence, depth maps and geometric descriptors offers promising applications for designing novel antibodies and vaccines.<\/jats:p>","DOI":"10.1371\/journal.pone.0335270","type":"journal-article","created":{"date-parts":[[2025,11,4]],"date-time":"2025-11-04T18:31:43Z","timestamp":1762281103000},"page":"e0335270","update-policy":"https:\/\/doi.org\/10.1371\/journal.pone.corrections_policy","source":"Crossref","is-referenced-by-count":0,"title":["Delineating SARS-CoV-2 spike protein and antibodies interaction interfaces via siamese neural networks: A geometric and image-based analysis"],"prefix":"10.1371","volume":"20","author":[{"given":"Gemma","family":"Loreti","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6171-5427","authenticated-orcid":true,"given":"Paola","family":"Vottero","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Elena Carlotta","family":"Olivetti","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Enrico","family":"Vezzetti","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jack","family":"Tuszynski","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4360-6905","authenticated-orcid":true,"given":"Federica","family":"Marcolin","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Maral","family":"Aminpour","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"340","published-online":{"date-parts":[[2025,11,4]]},"reference":[{"key":"pone.0335270.ref001","unstructured":"David L, Nelson DL, Cox MM, Stiedemann L, McGlynn ME, Fay MR. Lehninger principles of biochemistry. 2000."},{"volume-title":"Immunobiology: the immune system in health and disease","year":"2001","author":"CAJ Janeway","key":"pone.0335270.ref002"},{"issue":"1","key":"pone.0335270.ref003","doi-asserted-by":"crossref","first-page":"3","DOI":"10.1038\/s41580-021-00418-x","article-title":"Mechanisms of SARS-CoV-2 entry into cells","volume":"23","author":"CB Jackson","year":"2022","journal-title":"Nat Rev Mol Cell Biol."},{"issue":"1","key":"pone.0335270.ref004","doi-asserted-by":"crossref","first-page":"46","DOI":"10.1038\/nri.2017.106","article-title":"Beyond binding: antibody effector functions in infectious diseases","volume":"18","author":"LL Lu","year":"2018","journal-title":"Nat Rev Immunol."},{"key":"pone.0335270.ref005","doi-asserted-by":"crossref","first-page":"465","DOI":"10.1007\/978-3-030-41769-7_19","article-title":"Antigen-antibody complexes","volume":"94","author":"AB Kapingidza","year":"2020","journal-title":"Subcell Biochem."},{"issue":"2","key":"pone.0335270.ref006","doi-asserted-by":"crossref","first-page":"269","DOI":"10.1016\/0022-2836(82)90153-X","article-title":"A geometric approach to macromolecule-ligand interactions","volume":"161","author":"ID Kuntz","year":"1982","journal-title":"J Mol Biol."},{"issue":"1","key":"pone.0335270.ref007","doi-asserted-by":"crossref","first-page":"133","DOI":"10.1006\/jmbi.1997.1233","article-title":"Prediction of protein-protein interaction sites using patch analysis","volume":"272","author":"S Jones","year":"1997","journal-title":"J Mol Biol."},{"issue":"14","key":"pone.0335270.ref008","doi-asserted-by":"crossref","first-page":"5082","DOI":"10.1021\/jo052659z","article-title":"Shapes of antibody binding sites: qualitative and quantitative analyses based on a geomorphic classification scheme","volume":"71","author":"M Lee","year":"2006","journal-title":"J Org Chem."},{"issue":"04","key":"pone.0335270.ref009","doi-asserted-by":"crossref","first-page":"669","DOI":"10.1142\/S0218001493000339","article-title":"Signature verification using a \u201cSIAMESE\u201d time delay neural network","volume":"07","author":"J Bromley","year":"1993","journal-title":"Int J Patt Recogn Artif Intell."},{"key":"pone.0335270.ref010","unstructured":"Koch GR. Siamese neural networks for one-shot image recognition. 2015. https:\/\/api.semanticscholar.org\/CorpusID:13874643"},{"issue":"14","key":"pone.0335270.ref011","doi-asserted-by":"crossref","DOI":"10.1093\/bioinformatics\/btz328","article-title":"Multifaceted protein-protein interaction prediction based on Siamese residual RCNN","volume":"35","author":"M Chen","year":"2019","journal-title":"Bioinformatics."},{"issue":"24","key":"pone.0335270.ref012","doi-asserted-by":"crossref","first-page":"4771","DOI":"10.1093\/bioinformatics\/btab533","article-title":"Transfer learning via multi-scale convolutional neural layers for human-virus protein-protein interaction prediction","volume":"37","author":"X Yang","year":"2021","journal-title":"Bioinformatics."},{"issue":"6","key":"pone.0335270.ref013","doi-asserted-by":"crossref","DOI":"10.1093\/bib\/bbab255","article-title":"Improved prediction of protein-protein interaction using a hybrid of functional-link Siamese neural network and gradient boosting machines","volume":"22","author":"S Mahapatra","year":"2021","journal-title":"Brief Bioinform."},{"key":"pone.0335270.ref014","doi-asserted-by":"crossref","unstructured":"Lin T-Y, Dollar P, Girshick R, He K, Hariharan B, Belongie S. Feature pyramid networks for object detection. In: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2017. https:\/\/doi.org\/10.1109\/cvpr.2017.106","DOI":"10.1109\/CVPR.2017.106"},{"key":"pone.0335270.ref015","doi-asserted-by":"crossref","unstructured":"Wu J, Paquet E, Viktor HL, Michalowski W. Paying attention: using a siamese pyramid network for the prediction of protein-protein interactions with folding and self-binding primary sequences. In: 2021 International Joint Conference on Neural Networks (IJCNN). 2021. p. 1\u20138. https:\/\/doi.org\/10.1109\/ijcnn52387.2021.9534212","DOI":"10.1109\/IJCNN52387.2021.9534212"},{"issue":"1","key":"pone.0335270.ref016","doi-asserted-by":"crossref","first-page":"555","DOI":"10.1186\/s12864-022-08772-6","article-title":"DCSE: Double-Channel-Siamese-Ensemble model for protein protein interaction prediction","volume":"23","author":"W Chen","year":"2022","journal-title":"BMC Genomics."},{"issue":"8","key":"pone.0335270.ref017","doi-asserted-by":"crossref","first-page":"1144","DOI":"10.1093\/bioinformatics\/btv737","article-title":"DeNovo: virus-host sequence-based protein-protein interaction prediction","volume":"32","author":"F-E Eid","year":"2016","journal-title":"Bioinformatics."},{"key":"pone.0335270.ref018","article-title":"CodeTrans: towards cracking the language of silicone\u2019s code through self-supervised deep learning and high performance computing","author":"A Elnaggar","year":"2021","journal-title":"arXiv preprint"},{"issue":"9","key":"pone.0335270.ref019","doi-asserted-by":"crossref","first-page":"100551","DOI":"10.1016\/j.patter.2022.100551","article-title":"Accurate prediction of virus-host protein-protein interactions via a Siamese neural network using deep protein sequence embeddings","volume":"3","author":"S Madan","year":"2022","journal-title":"Patterns (N Y)."},{"issue":"2","key":"pone.0335270.ref020","doi-asserted-by":"crossref","first-page":"184","DOI":"10.1038\/s41592-019-0666-6","article-title":"Deciphering interaction fingerprints from protein molecular surfaces using geometric deep learning","volume":"17","author":"P Gainza","year":"2020","journal-title":"Nat Methods."},{"key":"pone.0335270.ref021","unstructured":"Gainza P, Sverrisson F, Boscaini D, Bronstein MM, Correia BE. Learning interaction patterns from surface representations of protein structure. 2019. https:\/\/api.semanticscholar.org\/CorpusID:231932111"},{"issue":"7959","key":"pone.0335270.ref022","doi-asserted-by":"crossref","first-page":"176","DOI":"10.1038\/s41586-023-05993-x","article-title":"De novo design of protein interactions with learned surface fingerprints","volume":"617","author":"P Gainza","year":"2023","journal-title":"Nature."},{"issue":"17","key":"pone.0335270.ref023","doi-asserted-by":"crossref","first-page":"2580","DOI":"10.1093\/bioinformatics\/btab154","article-title":"Protein interaction interface region prediction by geometric deep learning","volume":"37","author":"B Dai","year":"2021","journal-title":"Bioinformatics."},{"issue":"1","key":"pone.0335270.ref024","doi-asserted-by":"crossref","first-page":"1761","DOI":"10.1038\/s41598-020-80900-2","article-title":"Classification and prediction of protein-protein interaction interface using machine learning algorithm","volume":"11","author":"S Das","year":"2021","journal-title":"Sci Rep."},{"issue":"1","key":"pone.0335270.ref025","doi-asserted-by":"crossref","DOI":"10.1093\/bioinformatics\/btac797","article-title":"DockNet: high-throughput protein-protein interface contact prediction","volume":"39","author":"NP Williams","year":"2023","journal-title":"Bioinformatics."},{"issue":"1","key":"pone.0335270.ref026","doi-asserted-by":"crossref","first-page":"2175","DOI":"10.1038\/s41467-023-37701-8","article-title":"PeSTo: parameter-free geometric deep learning for accurate prediction of protein binding interfaces","volume":"14","author":"LF Krapp","year":"2023","journal-title":"Nat Commun."},{"issue":"7","key":"pone.0335270.ref027","first-page":"7471","article-title":"EGGNet, a generalizable geometric deep learning framework for protein complex pose scoring","volume":"9","author":"Z Wang","year":"2024","journal-title":"ACS Omega."},{"key":"pone.0335270.ref028","unstructured":"SAbDab: The Structural Antibody Database. [cited in 2023 Oct 1]. https:\/\/opig.stats.ox.ac.uk\/webapps\/sabdab-sabpred\/sabdab"},{"key":"pone.0335270.ref029","unstructured":"Protein Data Bank Structure: 1BVK. https:\/\/www.rcsb.org\/structure\/1BVK"},{"key":"pone.0335270.ref030","unstructured":"CoV-AbDab: The Coronavirus Antibody Database. https:\/\/opig.stats.ox.ac.uk\/webapps\/covabdab\/"},{"issue":"11","key":"pone.0335270.ref031","doi-asserted-by":"crossref","first-page":"1422","DOI":"10.1093\/bioinformatics\/btp163","article-title":"Biopython: freely available Python tools for computational molecular biology and bioinformatics","volume":"25","author":"PJA Cock","year":"2009","journal-title":"Bioinformatics."},{"issue":"9","key":"pone.0335270.ref032","article-title":"Quality improvement of surface triangular mesh using a modified Laplacian smoothing approach avoiding intersection","volume":"12","author":"T Liu","year":"2017","journal-title":"PLoS One."},{"key":"pone.0335270.ref033","doi-asserted-by":"crossref","DOI":"10.1093\/nar\/gkh381","article-title":"PDB2PQR: an automated pipeline for the setup of Poisson-Boltzmann electrostatics calculations","volume":"32","author":"TJ Dolinsky","year":"2004","journal-title":"Nucleic Acids Res."},{"issue":"1","key":"pone.0335270.ref034","doi-asserted-by":"crossref","first-page":"76","DOI":"10.1002\/bip.10207","article-title":"Gaussian docking functions","volume":"68","author":"MR McGann","year":"2003","journal-title":"Biopolymers."},{"issue":"2","key":"pone.0335270.ref035","doi-asserted-by":"crossref","first-page":"231","DOI":"10.1002\/bip.360330205","article-title":"Shape analysis of molecular surfaces","volume":"33","author":"BS Duncan","year":"1993","journal-title":"Biopolymers."},{"key":"pone.0335270.ref036","unstructured":"Liu T, Chen M, Lu B. Efficient and qualified mesh generation for gaussian molecular surface using piecewise trilinear polynomial approximation. 2016. https:\/\/arxiv.org\/abs\/1611.03474"},{"volume-title":"Introduction to linear algebra","year":"2009","author":"G Strang","key":"pone.0335270.ref037"},{"issue":"11","key":"pone.0335270.ref038","doi-asserted-by":"crossref","first-page":"1049","DOI":"10.1109\/42.816070","article-title":"Survey: interpolation methods in medical image processing","volume":"18","author":"TM Lehmann","year":"1999","journal-title":"IEEE Trans Med Imaging."},{"key":"pone.0335270.ref039","doi-asserted-by":"crossref","first-page":"108670","DOI":"10.1016\/j.jmgm.2023.108670","article-title":"Understanding the contagiousness of Covid-19 strains: a geometric approach","volume":"126","author":"P Vottero","year":"2024","journal-title":"J Mol Graph Model."},{"issue":"12","key":"pone.0335270.ref040","doi-asserted-by":"crossref","first-page":"13805","DOI":"10.1007\/s11042-016-3741-3","article-title":"Novel descriptors for geometrical 3D face analysis","volume":"76","author":"F Marcolin","year":"2016","journal-title":"Multimed Tools Appl."},{"key":"pone.0335270.ref041","doi-asserted-by":"crossref","first-page":"108670","DOI":"10.1016\/j.jmgm.2023.108670","article-title":"Understanding the contagiousness of Covid-19 strains: a geometric approach","volume":"126","author":"P Vottero","year":"2024","journal-title":"J Mol Graph Model."},{"key":"pone.0335270.ref042","article-title":"A new method for protein characterization and classification using geometrical features for 3D face analysis: an example of tubulin structures","author":"L Di Grazia","year":"2020","journal-title":"Proteins."},{"issue":"8","key":"pone.0335270.ref043","doi-asserted-by":"crossref","first-page":"557","DOI":"10.1016\/0262-8856(92)90076-F","article-title":"Surface shape and curvature scales","volume":"10","author":"JJ Koenderink","year":"1992","journal-title":"Image and Vision Computing."},{"key":"pone.0335270.ref044","doi-asserted-by":"crossref","first-page":"73","DOI":"10.1007\/978-1-0716-0826-5_3","article-title":"Siamese neural networks: an overview","volume":"2190","author":"D Chicco","year":"2021","journal-title":"Methods Mol Biol."},{"key":"pone.0335270.ref045","unstructured":"Python Documentation. https:\/\/docs.python.org\/3\/"},{"key":"pone.0335270.ref046","unstructured":"PyTorch. https:\/\/pytorch.org\/"},{"key":"pone.0335270.ref047","unstructured":"CUDA Documentation. [cited in 2023 Oct 1]. https:\/\/docs.nvidia.com\/cuda\/"},{"issue":"2","key":"pone.0335270.ref048","doi-asserted-by":"crossref","first-page":"298","DOI":"10.1093\/bioinformatics\/btv552","article-title":"ANARCI: antigen receptor numbering and receptor classification","volume":"32","author":"J Dunbar","year":"2016","journal-title":"Bioinformatics."},{"issue":"23","key":"pone.0335270.ref049","doi-asserted-by":"crossref","first-page":"3150","DOI":"10.1093\/bioinformatics\/bts565","article-title":"CD-HIT: accelerated for clustering the next-generation sequencing data","volume":"28","author":"L Fu","year":"2012","journal-title":"Bioinformatics."},{"issue":"6736","key":"pone.0335270.ref050","doi-asserted-by":"crossref","first-page":"850","DOI":"10.1126\/science.ads0018","article-title":"Simulating 500 million years of evolution with a language model","volume":"387","author":"T Hayes","year":"2025","journal-title":"Science."}],"container-title":["PLOS One"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0335270","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,11,4]],"date-time":"2025-11-04T18:31:51Z","timestamp":1762281111000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0335270"}},"subtitle":[],"editor":[{"given":"Yang","family":"Zhang","sequence":"first","affiliation":[],"role":[{"role":"editor","vocabulary":"crossref"}]}],"short-title":[],"issued":{"date-parts":[[2025,11,4]]},"references-count":50,"journal-issue":{"issue":"11","published-online":{"date-parts":[[2025,11,4]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pone.0335270","relation":{},"ISSN":["1932-6203"],"issn-type":[{"type":"electronic","value":"1932-6203"}],"subject":[],"published":{"date-parts":[[2025,11,4]]}}}