{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,8]],"date-time":"2026-05-08T00:47:42Z","timestamp":1778201262773,"version":"3.51.4"},"reference-count":42,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2024,10,23]],"date-time":"2024-10-23T00:00:00Z","timestamp":1729641600000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2024,10,23]],"date-time":"2024-10-23T00:00:00Z","timestamp":1729641600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/100010269","name":"Wellcome Trust","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100010269","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["J Cheminform"],"abstract":"<jats:sec>\n            <jats:title>Abstract<\/jats:title>\n            <jats:p>Neural processes (NPs) are models for meta-learning which output uncertainty estimates. So far, most studies of NPs have focused on low-dimensional datasets of highly-correlated tasks. While these homogeneous datasets are useful for benchmarking, they may not be representative of realistic transfer\u00a0learning. In particular, applications in scientific research may prove especially challenging due to the potential novelty of meta-testing tasks. Molecular property prediction is one such research area that is characterized by sparse datasets of many functions on a shared molecular space. In this paper, we study the application of graph NPs to molecular property prediction with DOCKSTRING, a diverse dataset of docking scores. Graph NPs show competitive performance in few-shot learning tasks relative to supervised learning baselines common in chemoinformatics, as well as alternative techniques for transfer learning and meta-learning. In order to increase meta-generalization to divergent test functions, we propose fine-tuning strategies that adapt the parameters of NPs. We find that adaptation can substantially increase NPs' regression performance while maintaining good calibration of uncertainty estimates. Finally, we present a Bayesian optimization experiment which showcases the potential advantages of NPs over Gaussian processes in iterative screening. Overall, our results suggest that NPs on molecular graphs hold great potential for molecular property prediction in the low-data setting.<\/jats:p>\n          <\/jats:sec>\n          <jats:sec>\n            <jats:title>Scientific contribution<\/jats:title>\n            <jats:p>Neural processes are a family of meta-learning algorithms which deal with data scarcity by transferring information across tasks and making probabilistic predictions. We evaluate their performance on regression and optimization molecular tasks using docking scores, finding them to outperform classical single-task and transfer-learning models. We examine the issue of generalization to divergent test tasks, which is a general concern of meta-learning algorithms in science, and propose strategies to alleviate it.<\/jats:p>\n          <\/jats:sec>","DOI":"10.1186\/s13321-024-00904-2","type":"journal-article","created":{"date-parts":[[2024,10,23]],"date-time":"2024-10-23T11:02:26Z","timestamp":1729681346000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":5,"title":["Graph neural processes for molecules: an evaluation on docking scores and strategies to improve generalization"],"prefix":"10.1186","volume":"16","author":[{"given":"Miguel","family":"Garc\u00eda-Orteg\u00f3n","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Srijit","family":"Seal","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Carl","family":"Rasmussen","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Andreas","family":"Bender","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Sergio","family":"Bacallado","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"297","published-online":{"date-parts":[[2024,10,23]]},"reference":[{"issue":"6","key":"904_CR1","doi-asserted-by":"publisher","first-page":"2848","DOI":"10.1021\/acs.jcim.0c00443","volume":"60","author":"BWJ Irwin","year":"2020","unstructured":"Irwin BWJ, Levell JR, Whitehead TM, Segall MD, Conduit GJ (2020) Practical applications of deep learning to impute heterogeneous drug discovery data. J Chem Inf Model. 60(6):2848\u20132857","journal-title":"J Chem Inf Model."},{"key":"904_CR2","unstructured":"ChEMBL (2023) ChEMBL Web Services. https:\/\/www.ebi.ac.uk\/chembl\/g\/#search_results\/all. Accessed 9 Aug 2023"},{"key":"904_CR3","unstructured":"Finn C, Abbeel P, Levine S (2017) Model-agnostic meta-learning for fast adaptation of deep networks. In: International Conference on Machine Learning. PMLR, pp 1126\u20131135"},{"key":"904_CR4","doi-asserted-by":"crossref","unstructured":"Nguyen CQ, Kreatsoulas C, Branson KM (2020) Meta-learning GNN initializations for low-resource molecular property prediction. In: ICML 2020 Workshop on Graph Representation Learning and Beyond (GRL+)","DOI":"10.26434\/chemrxiv.11981622.v1"},{"key":"904_CR5","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.ddtec.2020.11.009","volume":"37","author":"O Wieder","year":"2020","unstructured":"Wieder O, Kohlbacher S, Kuenemann M, Garon A, Ducrot P, Seidel T, Langer T (2020) A compact review of molecular property prediction with graph neural networks. Drug Discov Today: Technol 37:1\u201312. https:\/\/doi.org\/10.1016\/j.ddtec.2020.11.009","journal-title":"Drug Discov Today: Technol"},{"issue":"3","key":"904_CR6","doi-asserted-by":"publisher","first-page":"53","DOI":"10.1016\/j.slasd.2022.12.003","volume":"28","author":"A Liu","year":"2023","unstructured":"Liu A, Seal S, Yang H, Bender A (2023) Using chemical and biological data to predict drug toxicity. SLAS Discov 28(3):53\u201364. https:\/\/doi.org\/10.1016\/j.slasd.2022.12.003","journal-title":"SLAS Discov"},{"issue":"4","key":"904_CR7","doi-asserted-by":"publisher","first-page":"1040","DOI":"10.1016\/j.drudis.2020.11.037","volume":"26","author":"A Bender","year":"2021","unstructured":"Bender A, Cortes-Ciriano I (2021) Artificial intelligence in drug discovery: what is realistic, what are illusions? Part 2: a discussion of chemical and biological data. Drug Discov Today 26(4):1040\u20131052","journal-title":"Drug Discov Today"},{"key":"904_CR8","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1007\/978-1-0716-1787-8_1","volume":"2390","author":"M Thomas","year":"2022","unstructured":"Thomas M, Boardman A, Garcia-Ortegon M, Hongbin Y, Graaf C, Bender A (2022) Applications of artificial intelligence in drug design: opportunities and challenges. Methods Mol Biol. 2390:1\u201359","journal-title":"Methods Mol Biol."},{"issue":"7","key":"904_CR9","doi-asserted-by":"publisher","first-page":"2324977","DOI":"10.1001\/jamanetworkopen.2023.24977","volume":"6","author":"M Valerie Jentzsch","year":"2023","unstructured":"Valerie Jentzsch M (2023) Costs and causes of oncology drug attrition with the example of insulin-like growth factor-1 receptor. JAMA Netw Open 6(7):2324977","journal-title":"JAMA Netw Open"},{"key":"904_CR10","unstructured":"Garnelo M, Rosenbaum D, Maddison C, Ramalho T, Saxton D, Shanahan M, Teh YW, Rezende D, Eslami SA (2018) Conditional neural processes. In: International Conference on Machine Learning. PMLR, pp 1704\u20131713"},{"key":"904_CR11","unstructured":"Garnelo M, Schwarz J, Rosenbaum D, Viola F, Rezende DJ, Eslami SMA, Teh YW (2018) Neural Processes. In: ICML 2018 Workshop on Theoretical Foundations and Applications of Deep Generative Models"},{"key":"904_CR12","unstructured":"Antoniou A, Edwards H, Storkey A (2019) How to train your MAML. In: International Conference on Learning Representations"},{"key":"904_CR13","unstructured":"Chen W, Tripp A, Hern\u00e1ndez-Lobato JM (2023) Meta-learning adaptive deep kernel gaussian processes for molecular property prediction. In: International Conference on Learning Representations"},{"key":"904_CR14","unstructured":"Stanley M, Bronskill JF, Maziarz K, Misztela H, Lanini J, Segler M, Schneider N, Brockschmidt M (2021) FS-mol: a few-shot learning dataset of molecules. In: Thirty-fifth Conference on Neural Information Processing Systems Datasets and Benchmarks Track (Round 2)"},{"key":"904_CR15","unstructured":"Lee E, Yoo J, Lee H, Hong S (2022) MetaDTA: Meta-learning-based drug-target binding affinity prediction. In: ICLR 2022 Machine Learning for Drug Discovery Workshop"},{"key":"904_CR16","unstructured":"Garcia-Ortegon M, Bender A, Bacallado S (2022) Conditional neural processes for molecules. In: NeurIPS 2022 Workshop on Meta-Learning"},{"key":"904_CR17","unstructured":"Chan L, Verdonk M, Poelking C (2023) Embracing assay heterogeneity with neural processes for markedly improved bioactivity predictions"},{"issue":"1","key":"904_CR18","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s13321-022-00611-w","volume":"14","author":"M Walter","year":"2022","unstructured":"Walter M, Allen LN, Le\u00f3n A, Webb SJ, Gillet VJ (2022) Analysis of the benefits of imputation models over traditional QSAR models for toxicity prediction. J Cheminf. 14(1):1\u201327","journal-title":"J Cheminf."},{"issue":"12","key":"904_CR19","doi-asserted-by":"publisher","first-page":"3688","DOI":"10.1021\/acs.jcim.3c00132","volume":"63","author":"S Luukkonen","year":"2023","unstructured":"Luukkonen S, Meijer E, Tricarico GA, Hofmans J, Stouten PFW, Westen GJP, Lenselink EB (2023) Large-scale modeling of sparse protein kinase activity data. J Chem Inf Model. 63(12):3688\u20133696","journal-title":"J Chem Inf Model."},{"issue":"1","key":"904_CR20","doi-asserted-by":"publisher","first-page":"21","DOI":"10.1111\/j.1749-6632.2003.tb03158.x","volume":"994","author":"JES Wikberg","year":"2003","unstructured":"Wikberg JES, Mutulis F, Mutule I, Veiksina S, Lapinsh M, Petrovska R, Prusis P (2003) Melanocortin receptors: ligands and proteochemometrics modeling. Ann NY Acad Sci. 994(1):21\u201326","journal-title":"Ann NY Acad Sci."},{"key":"904_CR21","doi-asserted-by":"publisher","first-page":"89","DOI":"10.1016\/j.ddtec.2020.08.003","volume":"32","author":"BJ Bongers","year":"2019","unstructured":"Bongers BJ, IJzerman AP, Van Westen GJP (2019) Proteochemometrics: recent developments in bioactivity and selectivity modeling. Drug Discov Today Technol 32:89\u201398","journal-title":"Drug Discov Today Technol"},{"issue":"15","key":"904_CR22","doi-asserted-by":"publisher","first-page":"3486","DOI":"10.1021\/acs.jcim.1c01334","volume":"62","author":"M Garc\u00eda-Orteg\u00f3n","year":"2022","unstructured":"Garc\u00eda-Orteg\u00f3n M, Simm GNC, Tripp AJ, Hern\u00e1ndez-Lobato JM, Bender A, Bacallado S (2022) DOCKSTRING: easy molecular docking yields better benchmarks for ligand design. J Chem Inf Model. 62(15):3486\u20133502","journal-title":"J Chem Inf Model."},{"key":"904_CR23","doi-asserted-by":"publisher","DOI":"10.1021\/acs.jcim.4c00049","author":"GA Landrum","year":"2024","unstructured":"Landrum GA, Riniker S (2024) Combining IC50 or Ki values from different sources is a source of significant noise. J Chem Inf Model. https:\/\/doi.org\/10.1021\/acs.jcim.4c00049","journal-title":"J Chem Inf Model."},{"issue":"1","key":"904_CR24","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/1471-2105-12-257","volume":"12","author":"U Visser","year":"2011","unstructured":"Visser U, Abeyruwan S, Vempati U, Smith RP, Lemmon V, Sch\u00fcrer SC (2011) BioAssay Ontology (BAO): a semantic description of bioassays and high-throughput screening results. BMC Bioinf. 12(1):1\u201316. https:\/\/doi.org\/10.1186\/1471-2105-12-257","journal-title":"BMC Bioinf."},{"key":"904_CR25","unstructured":"Vinod R, Chen P-Y, Das P (2023) Reprogramming pretrained language models for protein sequence representation learning"},{"key":"904_CR26","unstructured":"Formont P, Jeannin H, Piantanida P, Ayed IB (2024) Is meta-training really necessary for molecular few-shot learning? arXiv. arXiv:2404.02314"},{"issue":"16","key":"904_CR27","doi-asserted-by":"publisher","first-page":"8749","DOI":"10.1021\/acs.jmedchem.9b00959","volume":"63","author":"Z Xiong","year":"2020","unstructured":"Xiong Z, Wang D, Liu X, Zhong F, Wan X, Li X, Li Z, Luo X, Chen K, Jiang H, Zheng M (2020) Pushing the boundaries of molecular representation for drug discovery with the graph attention mechanism. J Med Chem. 63(16):8749\u20138760","journal-title":"J Med Chem."},{"key":"904_CR28","unstructured":"Chung J, Gulcehre C, Cho K, Bengio Y (2014) Empirical evaluation of gated recurrent neural networks on sequence modeling. In: NeurIPS 2014 Deep Learning and Representation Learning Workshop"},{"key":"904_CR29","unstructured":"Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez AN, Kaiser \u0141, Polosukhin I (2017) Attention is all you need. In: Advances in Neural Information Processing Systems, vol 30"},{"issue":"1","key":"904_CR30","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s13321-017-0203-5","volume":"9","author":"J Sun","year":"2017","unstructured":"Sun J, Jeliazkova N, Chupakhin V, Golib-Dzib J-F, Engkvist O, Carlsson L, Wegner J, Ceulemans H, Georgiev I, Jeliazkov V, Kochev N, Ashby TJ, Chen H (2017) ExCAPE-DB: an integrated large scale dataset facilitating Big Data analysis in chemogenomics. J Cheminf. 9(1):1\u20139. https:\/\/doi.org\/10.1186\/s13321-017-0203-5","journal-title":"J Cheminf."},{"issue":"14","key":"904_CR31","doi-asserted-by":"publisher","first-page":"6582","DOI":"10.1021\/jm300687e","volume":"55","author":"MM Mysinger","year":"2012","unstructured":"Mysinger MM, Carchia M, Irwin John J, Shoichet BK (2012) Directory of useful decoys, enhanced (DUD-E): better ligands and decoys for better benchmarking. J Med Chem. 55(14):6582\u20136594. https:\/\/doi.org\/10.1021\/jm300687e","journal-title":"J Med Chem."},{"issue":"2","key":"904_CR32","doi-asserted-by":"publisher","first-page":"513","DOI":"10.1039\/C7SC02664A","volume":"9","author":"Z Wu","year":"2018","unstructured":"Wu Z, Ramsundar B, Feinberg EN, Gomes J, Geniesse C, Pappu AS, Leswing K, Pande V (2018) MoleculeNet: a benchmark for molecular machine learning. Chem Sci. 9(2):513\u2013530. https:\/\/doi.org\/10.1039\/C7SC02664A","journal-title":"Chem Sci."},{"issue":"4","key":"904_CR33","doi-asserted-by":"publisher","first-page":"783","DOI":"10.1021\/ci400084k","volume":"53","author":"RP Sheridan","year":"2013","unstructured":"Sheridan RP (2013) Time-split cross-validation as a method for estimating the goodness of prospective prediction. J Chem Inf Model. 53(4):783\u2013790","journal-title":"J Chem Inf Model."},{"issue":"8","key":"904_CR34","doi-asserted-by":"publisher","first-page":"2077","DOI":"10.1021\/acs.jcim.7b00166","volume":"57","author":"EJ Martin","year":"2017","unstructured":"Martin EJ, Polyakov VR, Tian L, Perez RC (2017) Profile-QSAR 2.0: kinase virtual screening accuracy comparable to four-concentration IC50s for realistically novel compounds. J Chem Inf Model. 57(8):2077\u20132088","journal-title":"J Chem Inf Model."},{"issue":"1","key":"904_CR35","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s13321-021-00576-2","volume":"13","author":"J Simm","year":"2021","unstructured":"Simm J, Humbeck L, Zalewski A, Sturm N, Heyndrickx W, Moreau Y, Beck B, Schuffenhauer A (2021) Splitting chemical structure data sets for federated privacy-preserving machine learning. J Cheminf. 13(1):1\u201314","journal-title":"J Cheminf."},{"issue":"2","key":"904_CR36","doi-asserted-by":"publisher","first-page":"107","DOI":"10.1021\/c160017a018","volume":"5","author":"HL Morgan","year":"1965","unstructured":"Morgan HL (1965) The generation of a unique machine description for chemical structures-a technique developed at chemical abstracts service. J Chem Doc. 5(2):107\u2013113","journal-title":"J Chem Doc."},{"key":"904_CR37","unstructured":"RDKit: RDKit: Open-source cheminformatics. https:\/\/www.rdkit.org"},{"issue":"1","key":"904_CR38","doi-asserted-by":"publisher","first-page":"31","DOI":"10.1021\/ci00057a005","volume":"28","author":"D Weininger","year":"1988","unstructured":"Weininger D (1988) SMILES, a chemical language and information system. 1. Introduction to methodology and encoding rules. J Chem Inf Comput Sci. 28(1):31\u201336","journal-title":"J Chem Inf Comput Sci."},{"key":"904_CR39","doi-asserted-by":"crossref","unstructured":"Garcia-Ortegon M, Seal S, Singh S, Bender A, Bacallado S (2024) Calibrated prediction of scarce adverse drug reaction labels with conditional neural processes. bioRxiv, 2024\u20130607598036.2024.06.07.598036","DOI":"10.1101\/2024.06.07.598036"},{"key":"904_CR40","doi-asserted-by":"publisher","first-page":"1256","DOI":"10.1038\/s42256-022-00580-7","volume":"4","author":"J Ross","year":"2022","unstructured":"Ross J, Belgodere B, Chenthamarakshan V, Padhi I, Mroueh Y, Das P (2022) Large-scale chemical language representations capture molecular structure and properties. Nat Mach Intell. 4:1256\u20131264. https:\/\/doi.org\/10.1038\/s42256-022-00580-7","journal-title":"Nat Mach Intell."},{"issue":"1","key":"904_CR41","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s12859-024-05847-x","volume":"25","author":"S Sadeghi","year":"2024","unstructured":"Sadeghi S, Bui A, Forooghi A, Lu J, Ngom A (2024) Can large language models understand molecules? BMC Bioinf. 25(1):1\u201317. https:\/\/doi.org\/10.1186\/s12859-024-05847-x","journal-title":"BMC Bioinf."},{"key":"904_CR42","unstructured":"Hu EJ, Shen Y, Wallis P, Allen-Zhu Z, Li Y, Wang S, Wang L, Chen W (2022) LoRA: Low-rank adaptation of large language models. In: International Conference on Learning Representations. https:\/\/openreview.net\/forum?id=nZeVKeeFYf9"}],"container-title":["Journal of Cheminformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s13321-024-00904-2.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1186\/s13321-024-00904-2\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s13321-024-00904-2.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,9,5]],"date-time":"2025-09-05T23:43:53Z","timestamp":1757115833000},"score":1,"resource":{"primary":{"URL":"https:\/\/jcheminf.biomedcentral.com\/articles\/10.1186\/s13321-024-00904-2"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,10,23]]},"references-count":42,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2024,12]]}},"alternative-id":["904"],"URL":"https:\/\/doi.org\/10.1186\/s13321-024-00904-2","relation":{},"ISSN":["1758-2946"],"issn-type":[{"value":"1758-2946","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,10,23]]},"assertion":[{"value":"29 February 2024","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"13 September 2024","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"23 October 2024","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"11 July 2025","order":4,"name":"change_date","label":"Change Date","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"Update","order":5,"name":"change_type","label":"Change Type","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"This article has been updated to amend the license information","order":6,"name":"change_details","label":"Change Details","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors have nothing to declare.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"115"}}