{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,22]],"date-time":"2026-07-22T02:01:10Z","timestamp":1784685670640,"version":"3.55.0"},"reference-count":48,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2026,5,26]],"date-time":"2026-05-26T00:00:00Z","timestamp":1779753600000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,7,22]],"date-time":"2026-07-22T00:00:00Z","timestamp":1784678400000},"content-version":"vor","delay-in-days":57,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["BMC Bioinformatics"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    The protein inverse folding problem, which is the task of designing an amino acid sequence that will fold into a specified backbone structure, represents a fundamental challenge in\n                    <jats:italic>de novo<\/jats:italic>\n                    protein design. Existing computational methods, including deep learning-based approaches, often fail to simultaneously optimize accuracy, stability, efficiency, and generalizability across diverse folds. Here, we present ProtSeqGen, a deep learning model that overcomes these limitations through a multi-stage graph-based framework. ProtSeqGen encodes protein structures as local geometric graphs, explicitly models residue-level interactions using a message-passing neural network, and predicts optimal amino acids with a multi-layer perceptron. When trained on CATH 4.2 dataset and evaluated on standard and challenging benchmarks, ProtSeqGen achieved superior sequence recovery compared to numerous state-of-the-art (SOTA) methods. It also generated accurate, designable sequences for nine topologically diverse proteins, demonstrating remarkable generalization capability. These results establish ProtSeqGen as a robust and scalable solution to the protein inverse folding problem, propelling\n                    <jats:italic>de novo<\/jats:italic>\n                    protein design with high structural precision.\n                  <\/jats:p>","DOI":"10.1186\/s12859-026-06482-4","type":"journal-article","created":{"date-parts":[[2026,5,26]],"date-time":"2026-05-26T06:44:12Z","timestamp":1779777852000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["ProtSeqGen: a novel deep learning model for protein sequence design"],"prefix":"10.1186","volume":"27","author":[{"given":"Qiang","family":"Gao","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhijin","family":"Li","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yang","family":"Deng","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhiwei","family":"Ji","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,5,26]]},"reference":[{"issue":"5","key":"6482_CR1","doi-asserted-by":"publisher","first-page":"bbad261","DOI":"10.1093\/bib\/bbad261","volume":"24","author":"S Xie","year":"2023","unstructured":"Xie S, Xie X, Zhao X, et al. HNSPPI: a hybrid computational model combing network and sequence information for predicting protein-protein interaction. Brief Bioinform. 2023;24(5):bbad261.","journal-title":"Brief Bioinform"},{"key":"6482_CR2","doi-asserted-by":"publisher","first-page":"3168","DOI":"10.1038\/s41467-021-23303-9","volume":"12","author":"V Gligorijevic","year":"2021","unstructured":"Gligorijevic V, Renfrew PD, Kosciolek T, et al. Structure-based protein function prediction using graph convolutional networks. Nat Commun. 2021;12:3168.","journal-title":"Nat Commun"},{"key":"6482_CR3","doi-asserted-by":"publisher","first-page":"497","DOI":"10.1016\/S1359-0278(98)00066-2","volume":"3","author":"M Gerstein","year":"1998","unstructured":"Gerstein M. How representative are the known structures of the proteins in a complete genome? a comprehensive structural census. Fold Des. 1998;3:497\u2013512.","journal-title":"Fold Des"},{"issue":"3","key":"6482_CR4","doi-asserted-by":"publisher","first-page":"bbae135","DOI":"10.1093\/bib\/bbae135","volume":"25","author":"J Mu","year":"2024","unstructured":"Mu J, Li Z, Zhang B, et al. Graphormer supervised de novo protein design method and function validation. Brief Bioinform. 2024;25(3):bbae135.","journal-title":"Brief Bioinform"},{"key":"6482_CR5","doi-asserted-by":"publisher","first-page":"526","DOI":"10.1016\/j.cell.2023.12.028","volume":"187","author":"T Kortemme","year":"2024","unstructured":"Kortemme T. De novo protein design-From new structures to programmable functions. Cell. 2024;187:526\u201344.","journal-title":"Cell"},{"key":"6482_CR6","doi-asserted-by":"publisher","first-page":"217","DOI":"10.1007\/978-1-4939-6637-0_10","volume":"1529","author":"P Xiong","year":"2017","unstructured":"Xiong P, Chen Q, Liu H. Computational protein design under a given backbone structure with the ABACUS statistical energy function. Methods Mol Biol. 2017;1529:217\u201326.","journal-title":"Methods Mol Biol"},{"key":"6482_CR7","doi-asserted-by":"publisher","first-page":"e2017228118","DOI":"10.1073\/pnas.2017228118","volume":"118","author":"C Norn","year":"2021","unstructured":"Norn C, Wicky BIM, Juergens D, et al. Protein sequence design by conformational landscape optimization. Proc Natl Acad Sci U S A. 2021;118:e2017228118.","journal-title":"Proc Natl Acad Sci U S A"},{"key":"6482_CR8","doi-asserted-by":"publisher","first-page":"639","DOI":"10.1038\/s41580-024-00718-y","volume":"25","author":"D Listov","year":"2024","unstructured":"Listov D, Goverde CA, Correia BE, et al. Opportunities and challenges in design and optimization of protein function. Nat Rev Mol Cell Biol. 2024;25:639\u201353.","journal-title":"Nat Rev Mol Cell Biol"},{"key":"6482_CR9","doi-asserted-by":"publisher","first-page":"108603","DOI":"10.1016\/j.biotechadv.2025.108603","volume":"82","author":"XC Cui","year":"2025","unstructured":"Cui XC, Zheng Y, Liu Y, et al. AI-driven de novo enzyme design: Strategies, applications, and future prospects. Biotechnol Adv. 2025;82:108603.","journal-title":"Biotechnol Adv"},{"key":"6482_CR10","doi-asserted-by":"publisher","first-page":"1364","DOI":"10.1126\/science.1089427","volume":"302","author":"B Kuhlman","year":"2003","unstructured":"Kuhlman B, Dantas G, Ireton GC, et al. Design of a novel globular protein fold with atomic-level accuracy. Science. 2003;302:1364\u20138.","journal-title":"Science"},{"key":"6482_CR11","doi-asserted-by":"publisher","first-page":"10637","DOI":"10.1021\/acs.jpca.0c07605","volume":"124","author":"D Mignon","year":"2020","unstructured":"Mignon D, Druart K, Michael E, et al. Physics-Based Computational Protein Design: An Update. J Phys Chem A. 2020;124:10637\u201348.","journal-title":"J Phys Chem A"},{"key":"6482_CR12","doi-asserted-by":"publisher","first-page":"5330","DOI":"10.1038\/ncomms6330","volume":"5","author":"P Xiong","year":"2014","unstructured":"Xiong P, Wang M, Zhou X, et al. Protein design with a comprehensive statistical energy function and boosted by experimental selection for foldability. Nat Commun. 2014;5:5330.","journal-title":"Nat Commun"},{"key":"6482_CR13","doi-asserted-by":"publisher","first-page":"487","DOI":"10.1186\/1471-2105-9-487","volume":"9","author":"M Peto","year":"2008","unstructured":"Peto M, Kloczkowski A, Honavar V, et al. Use of machine learning algorithms to classify binary protein sequences as highly-designable or poorly-designable. BMC Bioinformatics. 2008;9:487.","journal-title":"BMC Bioinformatics"},{"key":"6482_CR14","doi-asserted-by":"publisher","first-page":"2565","DOI":"10.1002\/prot.24620","volume":"82","author":"Z Li","year":"2014","unstructured":"Li Z, Yang Y, Faraggi E, et al. Direct prediction of profiles of sequences compatible with a protein structure by neural networks with fragment-based local and energy-based nonlocal profiles. Proteins. 2014;82:2565\u201373.","journal-title":"Proteins"},{"issue":"3","key":"6482_CR15","doi-asserted-by":"publisher","first-page":"bbac102","DOI":"10.1093\/bib\/bbac102","volume":"23","author":"W Ding","year":"2022","unstructured":"Ding W, Nakai K, Gong H. Protein design via deep learning. Brief Bioinform. 2022;23(3):bbac102.","journal-title":"Brief Bioinform"},{"key":"6482_CR16","doi-asserted-by":"publisher","first-page":"4075","DOI":"10.1093\/bioinformatics\/btab411","volume":"37","author":"W Ding","year":"2021","unstructured":"Ding W, Xu Q, Liu S, et al. SAMF: a self-adaptive protein modeling framework. Bioinformatics. 2021;37:4075\u201382.","journal-title":"Bioinformatics"},{"key":"6482_CR17","doi-asserted-by":"publisher","first-page":"819","DOI":"10.1002\/prot.25868","volume":"88","author":"Y Zhang","year":"2020","unstructured":"Zhang Y, Chen Y, Wang C, et al. ProDCoNN: Protein design using a convolutional neural network. Proteins. 2020;88:819\u201329.","journal-title":"Proteins"},{"key":"6482_CR18","doi-asserted-by":"publisher","first-page":"391","DOI":"10.1021\/acs.jcim.9b00438","volume":"60","author":"S Chen","year":"2020","unstructured":"Chen S, Sun Z, Lin L, et al. To Improve protein sequence profile prediction through image captioning on pairwise residue distance map. J Chem Inf Model. 2020;60:391\u20139.","journal-title":"J Chem Inf Model"},{"key":"6482_CR19","doi-asserted-by":"crossref","unstructured":"Tan C, Gao Z, Xia J et al. ICASSP. Global-Context Aware Generative Protein Design, 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (2023) 2023:1\u20135.","DOI":"10.1109\/ICASSP49357.2023.10095229"},{"key":"6482_CR20","unstructured":"Jing B, Eismann S, Suriana P et al. Learning from protein structure with geometric vector perceptrons, ICLR 2021 2021:1\u201318."},{"key":"6482_CR21","doi-asserted-by":"publisher","first-page":"49","DOI":"10.1126\/science.add2187","volume":"378","author":"J Dauparas","year":"2022","unstructured":"Dauparas J, Anishchenko I, Bennett N, et al. Robust deep learning-based protein sequence design using ProteinMPNN. Science. 2022;378:49\u201356.","journal-title":"Science"},{"key":"6482_CR22","unstructured":"Gao Z, Tan C, Li S, Pifold. Toward effective and efficient protein inverse folding, International Conference on Learning RepresentationsICLR (2023) 2023:1\u201314."},{"key":"6482_CR23","doi-asserted-by":"publisher","first-page":"e202411461","DOI":"10.1002\/anie.202411461","volume":"63","author":"J Liu","year":"2024","unstructured":"Liu J, Guo Z, You H, et al. All-Atom protein sequence design based on geometric deep learning. Angew Chem Int Ed Engl. 2024;63:e202411461.","journal-title":"Angew Chem Int Ed Engl"},{"issue":"12","key":"6482_CR24","doi-asserted-by":"publisher","first-page":"1034","DOI":"10.1038\/s44222-025-00349-8","volume":"3","author":"H Koh","year":"2025","unstructured":"Koh H, Zheng Y, Yang M, et al. AI-driven protein design. Nat reviews Bioeng. 2025;3(12):1034\u201356.","journal-title":"Nat reviews Bioeng"},{"key":"6482_CR25","doi-asserted-by":"publisher","first-page":"1166","DOI":"10.1038\/s44320-025-00119-z","volume":"21","author":"MA Jendrusch","year":"2025","unstructured":"Jendrusch MA, Yang ALJ, Cacace E, et al. AlphaDesign: a De Novo protein design framework based on AlphaFold. Mol Syst Biol. 2025;21:1166\u201389.","journal-title":"Mol Syst Biol"},{"key":"6482_CR26","doi-asserted-by":"publisher","first-page":"897","DOI":"10.1038\/s41586-023-07004-5","volume":"626","author":"G Tesei","year":"2024","unstructured":"Tesei G, Trolle AI, Jonsson N, et al. Conformational ensembles of the human intrinsically disordered proteome. Nature. 2024;626:897\u2013904.","journal-title":"Nature"},{"issue":"22","key":"6482_CR27","doi-asserted-by":"publisher","first-page":"14050","DOI":"10.3390\/ijms232214050","volume":"23","author":"R Trivedi","year":"2022","unstructured":"Trivedi R, Nagarajaram HA. Intrinsically disordered proteins: an overview. Int J Mol Sci. 2022;23(22):14050.","journal-title":"Int J Mol Sci"},{"key":"6482_CR28","doi-asserted-by":"publisher","first-page":"1558","DOI":"10.1002\/prot.26533","volume":"91","author":"A Kryshtafovych","year":"2023","unstructured":"Kryshtafovych A, Rigden DJ. To split or not to split: CASP15 targets and their processing into tertiary structure evaluation units. Proteins. 2023;91:1558\u201370.","journal-title":"Proteins"},{"issue":"Suppl 1","key":"6482_CR29","doi-asserted-by":"publisher","first-page":"51","DOI":"10.1002\/prot.24973","volume":"84","author":"LN Kinch","year":"2016","unstructured":"Kinch LN, Li W, Monastyrskyy B, et al. Evaluation of free modeling targets in CASP11 and ROLL. Proteins. 2016;84(Suppl 1):51\u201366.","journal-title":"Proteins"},{"key":"6482_CR30","first-page":"28337","volume":"38","author":"M Liu","year":"2026","unstructured":"Liu M, Cheng X, Gao Z, et al. ProtInvTree: deliberate protein inverse folding with reward-guided tree search. Adv Neural Inf Process Syst. 2026;38:28337\u201360.","journal-title":"Adv Neural Inf Process Syst"},{"key":"6482_CR31","doi-asserted-by":"publisher","first-page":"9842315","DOI":"10.34133\/2022\/9842315","volume":"2022","author":"AE Chu","year":"2022","unstructured":"Chu AE, Fernandez D, Liu J, et al. De Novo design of a highly stable ovoid tim barrel: unlocking pocket shape towards functional design. Biodes Res. 2022;2022:9842315.","journal-title":"Biodes Res"},{"key":"6482_CR32","doi-asserted-by":"publisher","first-page":"8","DOI":"10.1038\/s41467-025-66526-w","volume":"17","author":"Y Cho","year":"2025","unstructured":"Cho Y, Dauparas J, Tsuboyama K, et al. Stable de novo protein design via joint conformational landscape and sequence optimization. Nat Commun. 2025;17:8.","journal-title":"Nat Commun"},{"key":"6482_CR33","unstructured":"Cai C, Hy T, Yu R et al. On the connection between mpnn and graph transformer. In: International conference on machine learning (ICML2023) 2023:3408\u20133430."},{"key":"6482_CR34","unstructured":"Hwang E, Thost V, Dasgupta S et al. IoG. An analysis of virtual nodes in graph neural networks for link prediction. In: The first learning on graphs conference. 2022:1\u201311."},{"key":"6482_CR35","unstructured":"Ingraham J, Garg V, Barzilay R et al. NeurIPS. Generative models for graph-based protein design. In: 33rd Conference on Neural Information Processing Systems. 2019:1\u201313."},{"key":"6482_CR36","unstructured":"Jing B, Eismann S, Suriana P et al. ICLR. Learning from protein structure with geometric vector perceptrons. In: International Conference on Learning Representations. 2021:1\u201318."},{"key":"6482_CR37","doi-asserted-by":"publisher","first-page":"657","DOI":"10.1107\/S2052252517008466","volume":"4","author":"J Pereira","year":"2017","unstructured":"Pereira J, Lamzin VS. A distance geometry-based description and validation of protein main-chain conformation. IUCrJ. 2017;4:657\u201370.","journal-title":"IUCrJ"},{"key":"6482_CR38","doi-asserted-by":"publisher","first-page":"583","DOI":"10.1038\/s41586-021-03819-2","volume":"596","author":"J Jumper","year":"2021","unstructured":"Jumper J, Evans R, Pritzel A, et al. Highly accurate protein structure prediction with AlphaFold. Nature. 2021;596:583\u20139.","journal-title":"Nature"},{"key":"6482_CR39","doi-asserted-by":"publisher","first-page":"1496","DOI":"10.1073\/pnas.1914677117","volume":"117","author":"J Yang","year":"2020","unstructured":"Yang J, Anishchenko I, Park H, et al. Improved protein structure prediction using predicted interresidue orientations. Proc Natl Acad Sci U S A. 2020;117:1496\u2013503.","journal-title":"Proc Natl Acad Sci U S A"},{"key":"6482_CR40","doi-asserted-by":"publisher","first-page":"33953","DOI":"10.1038\/s41598-025-10940-z","volume":"15","author":"N Manjegowda","year":"2025","unstructured":"Manjegowda N, Rao M. Path-based evaluation of deep learning models for solving inverse kinematics in a revolute-prismatic robot. Sci Rep. 2025;15:33953.","journal-title":"Sci Rep"},{"key":"6482_CR41","doi-asserted-by":"crossref","unstructured":"Kalra D, Barkeshli M. NeurIPS. Why Warmup the Learning Rate? Underlying Mechanisms and Improvements. In: 38th Conference on Neural Information Processing Systems. 2024:1\u201342.","DOI":"10.52202\/079017-3549"},{"key":"6482_CR42","doi-asserted-by":"publisher","first-page":"110751","DOI":"10.1016\/j.enzmictec.2025.110751","volume":"192","author":"JJ Juo","year":"2025","unstructured":"Juo JJ, Ng IS. De Novo enzyme design of a minimal and soluble carbonic anhydrase from Mesorhizobium loti assisted by molecular dynamics simulation. Enzyme Microb Technol. 2025;192:110751.","journal-title":"Enzyme Microb Technol"},{"key":"6482_CR43","doi-asserted-by":"publisher","first-page":"e2412926","DOI":"10.1002\/advs.202412926","volume":"12","author":"Z Luo","year":"2025","unstructured":"Luo Z, Geng A, Wei L, et al. CPL-Diff: A diffusion model for De Novo design of functional peptide sequences with fixed length. Adv Sci (Weinh). 2025;12:e2412926.","journal-title":"Adv Sci (Weinh)"},{"key":"6482_CR44","doi-asserted-by":"publisher","first-page":"629","DOI":"10.1002\/prot.25489","volume":"86","author":"J O\u2019Connell","year":"2018","unstructured":"O\u2019Connell J, Li Z, Hanson J, et al. SPIN2: Predicting sequence profiles from protein structures using deep neural networks. Proteins. 2018;86:629\u201333.","journal-title":"Proteins"},{"key":"6482_CR45","doi-asserted-by":"publisher","first-page":"6349","DOI":"10.1038\/s41598-018-24760-x","volume":"8","author":"J Wang","year":"2018","unstructured":"Wang J, Cao H, Zhang JZH, et al. Computational protein design with deep learning neural networks. Sci Rep. 2018;8:6349.","journal-title":"Sci Rep"},{"key":"6482_CR46","doi-asserted-by":"publisher","first-page":"1245","DOI":"10.1021\/acs.jcim.0c00043","volume":"60","author":"Y Qi","year":"2020","unstructured":"Qi Y, Zhang JZH, DenseCPD. Improving the accuracy of neural-network-based computational protein sequence design with DenseNet. J Chem Inf Model. 2020;60:1245\u201352.","journal-title":"J Chem Inf Model"},{"key":"6482_CR47","first-page":"1","volume":"2023","author":"D Akpinaroglu","year":"2023","unstructured":"Akpinaroglu D, Seki K, Zhu E, et al. Frame2seq: structure-conditioned masked language modeling for protein sequence design. Mach Learn Struct Biology Workshop NeurIPS. 2023;2023:1\u201314.","journal-title":"Mach Learn Struct Biology Workshop NeurIPS"},{"key":"6482_CR48","doi-asserted-by":"publisher","first-page":"D280","DOI":"10.1093\/nar\/gky1097","volume":"47","author":"I Sillitoe","year":"2019","unstructured":"Sillitoe I, Dawson N, Lewis TE, et al. CATH: expanding the horizons of structure-based functional annotations for genome sequences. Nucleic Acids Res. 2019;47:D280\u20134.","journal-title":"Nucleic Acids Res"}],"container-title":["BMC Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/article\/10.1186\/s12859-026-06482-4","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s12859-026-06482-4.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s12859-026-06482-4.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,22]],"date-time":"2026-07-22T01:05:12Z","timestamp":1784682312000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1186\/s12859-026-06482-4"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,5,26]]},"references-count":48,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2026,12]]}},"alternative-id":["6482"],"URL":"https:\/\/doi.org\/10.1186\/s12859-026-06482-4","relation":{},"ISSN":["1471-2105"],"issn-type":[{"value":"1471-2105","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,5,26]]},"assertion":[{"value":"19 March 2026","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"8 May 2026","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"26 May 2026","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"Not applicable.","order":1,"name":"Ethics","label":"Ethics approval and consent to participate","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Not applicable.","order":2,"name":"Ethics","label":"Consent for publication","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.","order":3,"name":"Ethics","label":"Competing interests","group":{"name":"EthicsHeading","label":"Declarations"}}],"article-number":"156"}}