{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,3]],"date-time":"2026-08-03T20:01:54Z","timestamp":1785787314264,"version":"3.56.0"},"reference-count":65,"publisher":"National Academy of Sciences","issue":"32","license":[{"start":{"date-parts":[[2026,8,3]],"date-time":"2026-08-03T00:00:00Z","timestamp":1785715200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100003977","name":"Israel Science Foundation","doi-asserted-by":"publisher","award":["1764\/21"],"award-info":[{"award-number":["1764\/21"]}],"id":[{"id":"10.13039\/501100003977","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100000854","name":"Human Frontier Science Program","doi-asserted-by":"publisher","award":["RGEC29\/2025."],"award-info":[{"award-number":["RGEC29\/2025."]}],"id":[{"id":"10.13039\/501100000854","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.pnas.org"],"crossmark-restriction":true},"short-container-title":["Proc. Natl. Acad. Sci. U.S.A."],"published-print":{"date-parts":[[2026,8,11]]},"abstract":"<jats:p>Establishing a coherent mapping of the relationships among all known proteins is crucial for elucidating processes of protein emergence and evolution. Yet the capacity to fully capture relationships of protein similarity is complicated by the nonstraightforward interplay between sequence and structure; indeed, proteins with unrelated sequences can adopt similar structures, and, conversely, proteins with similar or identical sequences can manifest radically different structures. Here, we introduce Contrastive Learning Sequence\u2013Structure (CLSS), a contrastive protein language model (PLM) trained to coembed sequence and structure information in a self-supervised manner, facilitating a holistic representation of protein relatedness. CLSS represents the structures and sequences of full domains and domain subsequences as vectors in the same high-dimensional latent space. We show that this approach yields meaningful shared representations, which recapitulate the extensive structure- and sequence-based knowledge encoded in human-curated hierarchical protein classification systems (ECOD and CATH). Moreover, the representations generated by CLSS outperform those generated by alternative state-of-the-art PLMs in downstream classification tasks. Notably, we show that even the far larger space of domain subsequences is successfully coembedded, establishing a PLM tailored to these evolutionarily meaningful objects. CLSS embeddings produce informative representations of the protein universe without further downstream processing, as we demonstrate by analyzing preferential associations between protein architectures and ligand types across protein space.<\/jats:p>","DOI":"10.1073\/pnas.2532702123","type":"journal-article","created":{"date-parts":[[2026,8,3]],"date-time":"2026-08-03T19:06:16Z","timestamp":1785783976000},"update-policy":"https:\/\/doi.org\/10.1073\/pnas.cm10313","source":"Crossref","is-referenced-by-count":0,"title":["Contrastive learning unites sequence and structure in a global representation of protein space"],"prefix":"10.1073","volume":"123","author":[{"given":"Guy","family":"Yanai","sequence":"first","affiliation":[{"id":[{"id":"https:\/\/ror.org\/02f009v59","id-type":"ROR","asserted-by":"publisher"}],"name":"Department of Computer Science, University of Haifa","place":["Haifa, Israel"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2492-047X","authenticated-orcid":false,"given":"Gabriel","family":"Axel","sequence":"additional","affiliation":[{"id":[{"id":"https:\/\/ror.org\/04mhzgx49","id-type":"ROR","asserted-by":"publisher"}],"name":"Department of Biochemistry and Molecular Biology, School of Neurobiology, Biochemistry and Biophysics, George S. Wise Faculty of Life Sciences, Tel Aviv University","place":["Ramat Aviv, Israel"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1773-0611","authenticated-orcid":false,"given":"Liam M.","family":"Longo","sequence":"additional","affiliation":[{"id":[{"id":"https:\/\/ror.org\/05dqf9946","id-type":"ROR","asserted-by":"publisher"}],"name":"Earth-Life Science Institute, Institute of Science Tokyo","place":["Tokyo, Japan"]},{"id":[{"id":"https:\/\/ror.org\/04yhya597","id-type":"ROR","asserted-by":"publisher"}],"name":"Blue Marble Space Institute of Science","place":["Seattle"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6901-832X","authenticated-orcid":false,"given":"Nir","family":"Ben-Tal","sequence":"additional","affiliation":[{"id":[{"id":"https:\/\/ror.org\/04mhzgx49","id-type":"ROR","asserted-by":"publisher"}],"name":"Department of Biochemistry and Molecular Biology, School of Neurobiology, Biochemistry and Biophysics, George S. Wise Faculty of Life Sciences, Tel Aviv University","place":["Ramat Aviv, Israel"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8523-1614","authenticated-orcid":false,"given":"Rachel","family":"Kolodny","sequence":"additional","affiliation":[{"id":[{"id":"https:\/\/ror.org\/02f009v59","id-type":"ROR","asserted-by":"publisher"}],"name":"Department of Computer Science, University of Haifa","place":["Haifa, Israel"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"341","published-online":{"date-parts":[[2026,8,3]]},"reference":[{"key":"e_1_3_4_1_2","doi-asserted-by":"publisher","DOI":"10.1126\/science.273.5275.595"},{"key":"e_1_3_4_2_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2628030100"},{"key":"e_1_3_4_3_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1403395111"},{"key":"e_1_3_4_4_2","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gkaa913"},{"key":"e_1_3_4_5_2","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/28.1.257"},{"key":"e_1_3_4_6_2","doi-asserted-by":"publisher","DOI":"10.1016\/S0969-2126(97)00260-8"},{"key":"e_1_3_4_7_2","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pcbi.1003926"},{"key":"e_1_3_4_8_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1707642114"},{"key":"e_1_3_4_9_2","doi-asserted-by":"publisher","DOI":"10.1093\/molbev\/msab017"},{"key":"e_1_3_4_10_2","doi-asserted-by":"publisher","DOI":"10.1002\/pro.4407"},{"key":"e_1_3_4_11_2","doi-asserted-by":"publisher","DOI":"10.7554\/eLife.64415"},{"key":"e_1_3_4_12_2","doi-asserted-by":"publisher","DOI":"10.7554\/eLife.09410"},{"key":"e_1_3_4_13_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1911742117"},{"key":"e_1_3_4_14_2","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pcbi.1004466"},{"key":"e_1_3_4_15_2","doi-asserted-by":"publisher","DOI":"10.1002\/pro.297"},{"key":"e_1_3_4_16_2","first-page":"11569","article-title":"Representation of the protein universe using classifications, maps and networks","volume":"111","author":"Ben-Tal N.","year":"2014","unstructured":"N. Ben-Tal, R. Kolodny, Representation of the protein universe using classifications, maps and networks. Proc. Natl. Acad. Sci. U.S.A. 111, 11569\u201311570 (2014).","journal-title":"Proc. Natl. Acad. Sci. U.S.A."},{"key":"e_1_3_4_17_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1102727108"},{"key":"e_1_3_4_18_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2636460100"},{"key":"e_1_3_4_19_2","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pcbi.1006969"},{"key":"e_1_3_4_20_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.0914097107"},{"key":"e_1_3_4_21_2","doi-asserted-by":"publisher","DOI":"10.1093\/bioinformatics\/bty585"},{"key":"e_1_3_4_22_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.cels.2022.01.003"},{"key":"e_1_3_4_23_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.sbi.2025.102986"},{"key":"e_1_3_4_24_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.sbi.2025.103004"},{"key":"e_1_3_4_25_2","doi-asserted-by":"publisher","DOI":"10.1126\/science.ads0018"},{"key":"e_1_3_4_26_2","doi-asserted-by":"crossref","unstructured":"J. Su C. Han Y. Zhou Saprot: Protein language modeling with structure-aware vocabulary. bioRxiv [Preprint] (2023). https:\/\/doi.org\/10.1101\/2023.10.01.560349 (Accessed 19 April 2024).","DOI":"10.1101\/2023.10.01.560349"},{"key":"e_1_3_4_27_2","unstructured":"K. Fl\u00f6ge OneProt: Towards multi-modal protein foundation models. arXiv [Preprint] (2024). https:\/\/doi.org\/10.48550\/arXiv.2411.04863 (Accessed 18 October 2025)."},{"key":"e_1_3_4_28_2","doi-asserted-by":"publisher","DOI":"10.1002\/advs.202404212"},{"key":"e_1_3_4_29_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.jmb.2025.168940"},{"key":"e_1_3_4_30_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-025-02836-0"},{"key":"e_1_3_4_31_2","doi-asserted-by":"publisher","DOI":"10.1201\/9781315113876"},{"key":"e_1_3_4_32_2","doi-asserted-by":"publisher","DOI":"10.1021\/acs.jcim.4c00689"},{"key":"e_1_3_4_33_2","unstructured":"Z. Zhang Joint representation learning on protein sequences and structures. arXiv [Preprint] (2023). https:\/\/doi.org\/10.48550\/arXiv.2303.06275 (Accessed 18 October 2023)."},{"key":"e_1_3_4_34_2","unstructured":"D. Chen Endowing protein language models with structural knowledge. arXiv [Preprint] (2024). https:\/\/doi.org\/10.48550\/arXiv.2401.14819 (Accessed 26 January 2024)."},{"key":"e_1_3_4_35_2","doi-asserted-by":"publisher","DOI":"10.1002\/prot.21770"},{"key":"e_1_3_4_36_2","doi-asserted-by":"publisher","DOI":"10.1002\/prot.26753"},{"key":"e_1_3_4_37_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.str.2015.08.018"},{"key":"e_1_3_4_38_2","unstructured":"A. Radford Learning transferable visual models from natural language supervision. arXiv [Preprint] (2021). https:\/\/doi.org\/10.48550\/arXiv.2103.00020 (Accessed 26 February 2021)."},{"key":"e_1_3_4_39_2","doi-asserted-by":"publisher","DOI":"10.1093\/nargab\/lqae150"},{"key":"e_1_3_4_40_2","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2539"},{"key":"e_1_3_4_41_2","doi-asserted-by":"crossref","unstructured":"X. Zhai LiT: Zero-shot transfer with locked-image text tuning. arXiv [Preprint] (2022). https:\/\/doi.org\/10.48550\/arXiv.2111.07991 (Accessed 22 June 2022).","DOI":"10.1109\/CVPR52688.2022.01759"},{"key":"e_1_3_4_42_2","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pcbi.1011586"},{"key":"e_1_3_4_43_2","first-page":"2579","article-title":"Visualizing data using t-SNE","volume":"9","author":"Van der Maaten L.","year":"2008","unstructured":"L. Van der Maaten, G. Hinton, Visualizing data using t-SNE. J. Mach. Learn. Res. 9, 2579\u201326050 (2008).","journal-title":"J. Mach. Learn. Res."},{"key":"e_1_3_4_44_2","unstructured":"M. Huh \u201cPosition: The platonic representation hypothesis\u201d in Forty-first International Conference on Machine Learning (2024)."},{"key":"e_1_3_4_45_2","unstructured":"Z. Zhang Structure-informed protein language model. arXiv [Preprint] (2024). https:\/\/doi.org\/10.48550\/arXiv.2402.05856 (Accessed 7 February 2024)."},{"key":"e_1_3_4_46_2","doi-asserted-by":"publisher","DOI":"10.1093\/nargab\/lqac043"},{"key":"e_1_3_4_47_2","doi-asserted-by":"publisher","DOI":"10.1002\/prot.26753"},{"key":"e_1_3_4_48_2","doi-asserted-by":"crossref","unstructured":"T. Corlett The history of enzyme evolution embedded in metabolism. bioRxiv [Preprint] (2025). https:\/\/doi.org\/10.1101\/2025.07.16.665256 (Accessed 17 March 2026).","DOI":"10.1101\/2025.07.16.665256"},{"key":"e_1_3_4_49_2","doi-asserted-by":"publisher","DOI":"10.1016\/S0014-5793(01)02236-0"},{"key":"e_1_3_4_50_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.jmb.2020.166788"},{"key":"e_1_3_4_51_2","doi-asserted-by":"publisher","DOI":"10.1006\/jmbi.2001.5227"},{"key":"e_1_3_4_52_2","doi-asserted-by":"publisher","DOI":"10.1038\/273443a0"},{"key":"e_1_3_4_53_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.jmb.2006.05.060"},{"key":"e_1_3_4_54_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2220778120"},{"key":"e_1_3_4_55_2","unstructured":"P. Hermosilla T. Ropinski Contrastive representation learning for 3D protein structures. arXiv [Preprint] (2022). https:\/\/doi.org\/10.48550\/arXiv.2205.15675 (Accessed 31 May 2022)."},{"key":"e_1_3_4_56_2","unstructured":"Z. Zhang Protein representation learning by geometric structure pretraining. arXiv [Preprint] (2022). https:\/\/doi.org\/10.48550\/arXiv.2203.06125 (Accessed 31 May 2022)."},{"key":"e_1_3_4_57_2","doi-asserted-by":"publisher","DOI":"10.1006\/jmbi.1999.2893"},{"key":"e_1_3_4_58_2","doi-asserted-by":"publisher","DOI":"10.1126\/science.1853201"},{"key":"e_1_3_4_59_2","doi-asserted-by":"publisher","DOI":"10.1038\/358086a0"},{"key":"e_1_3_4_60_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.sbi.2008.02.004"},{"key":"e_1_3_4_61_2","doi-asserted-by":"crossref","unstructured":"Y. Akiyama Scaling down protein language modeling with MSA Pairformer. bioRxiv [Preprint] (2025). https:\/\/doi.org\/10.1101\/2025.08.02.668173 (Accessed 3 August 2025).","DOI":"10.1101\/2025.08.02.668173"},{"key":"e_1_3_4_62_2","doi-asserted-by":"crossref","unstructured":"A. X. Lu Tokenized and continuous embedding compressions of protein sequence and structure. bioRxiv [Preprint] (2024). https:\/\/doi.org\/10.1101\/2024.08.06.606920 (Accessed 3 November 2024).","DOI":"10.1101\/2024.08.06.606920"},{"key":"e_1_3_4_63_2","article-title":"Orthogonal range searching","author":"Demaine E.","year":"2008","unstructured":"E. Demaine, \u201cOrthogonal range searching\u201d in Computational Geometry: Algorithms and Applications (Springer, 2008).","journal-title":"Computational Geometry: Algorithms and Applications"},{"key":"e_1_3_4_64_2","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1992.4.1.1"},{"key":"e_1_3_4_65_2","doi-asserted-by":"crossref","unstructured":"G. Yanai Data from \u201cCLSS: Contrastive learning unites sequence and structure in a global representation of protein space.\u201d GitHub. https:\/\/github.com\/guyyanai\/CLSS. Deposited 22 May 2026.","DOI":"10.1101\/2025.09.05.674454"}],"container-title":["Proceedings of the National Academy of Sciences"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/pnas.org\/doi\/pdf\/10.1073\/pnas.2532702123","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,3]],"date-time":"2026-08-03T19:07:18Z","timestamp":1785784038000},"score":1,"resource":{"primary":{"URL":"https:\/\/pnas.org\/doi\/10.1073\/pnas.2532702123"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,3]]},"references-count":65,"journal-issue":{"issue":"32","published-print":{"date-parts":[[2026,8,11]]}},"alternative-id":["10.1073\/pnas.2532702123"],"URL":"https:\/\/doi.org\/10.1073\/pnas.2532702123","relation":{},"ISSN":["0027-8424","1091-6490"],"issn-type":[{"value":"0027-8424","type":"print"},{"value":"1091-6490","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,3]]},"assertion":[{"value":"2025-11-17","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-06-16","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-08-03","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e2532702123"}}