{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,12]],"date-time":"2026-10-12T12:49:30Z","timestamp":1791809370773,"version":"4.3.4"},"reference-count":51,"publisher":"American Chemical Society (ACS)","issue":"19","license":[{"start":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T00:00:00Z","timestamp":1790640000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T00:00:00Z","timestamp":1790640000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"},{"start":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T00:00:00Z","timestamp":1790640000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-045"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["22473009"],"award-info":[{"award-number":["22473009"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Beijing Natural Science Foundation","award":["L259004"],"award-info":[{"award-number":["L259004"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,10,12]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>With the advancement of computer-aided retrobiosynthesis, numerous biosynthetic pathways have been predicted, exceeding the capacity of experimental validation. Effective classifiers are needed to identify feasible reactions. In this study, we collected 75,864 feasible reactions from biocatalysis databases and generated an equal number of infeasible reactions based on reaction rules. Following atom mapping focused on reaction centers and systematic reaction preprocessing, three datasets for training were constructed: the \u201cstereo\u201d dataset, which retained reaction stereochemical information; the \u201cnon-stereo\u201d dataset, which was a stereochemistry-agnostic version of the \u201cstereo\u201d dataset; and the \u201cmixed\u201d dataset, which comprised both. We established a total of 22 individual enzymatic reaction feasibility classification models, which include: eXtreme Gradient Boosting (XGBoost) and Deep Neural Network (DNN) models utilizing Reaction Fingerprints (RXNFP), Differential Reaction Fingerprint (DRFP), and our constructed Combined ECFP4 Reaction Fingerprints (c_ECFP4) for reaction representation, and Transformer models and fine-tuned ChemBERTa-77M-MLM (ChemMLM) models using reaction SMILES strings as the direct input. The results indicate that models utilizing the c_ECFP4 representation achieved the highest predictive performance, which effectively captured underlying enzymatic reaction mechanisms. Among them, Model 1A-M (based on XGBoost and \u201cmixed\u201d dataset) was identified as the optimal individual model, achieving Matthews Correlation Coefficient (MCC) values of 0.865 and 0.853 and Area Under Curve (AUC) values of 0.981 and 0.980 on the \u201cstereo\u201d and \u201cnon-stereo\u201d test sets, respectively. Furthermore, a consensus model enzymatic reaction feasibility classification (ERFC) integrating four reaction representations further improved predictive performance, achieving MCC values of 0.894 and 0.886 and an AUC of 0.986 on both test sets. Moreover, both models (Model 1A-M and ERFC) successfully validated a five-step biosynthetic pathway, demonstrating higher prediction accuracy than the previously reported DeepRFC and DORA-XGB models in identifying feasible reactions. All data, the individual Model 1A-M, and the consensus model ERFC are openly available, offering a reliable and flexible framework for predicting enzymatic reaction feasibility in the presence or absence of stereochemical information.<\/jats:p>","DOI":"10.1021\/acs.jcim.6c02294","type":"journal-article","created":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T08:02:15Z","timestamp":1790668935000},"page":"12613-12623","source":"Crossref","is-referenced-by-count":0,"title":["Enzymatic Reaction\nFeasibility Classification Using\nMachine Learning Methods"],"prefix":"10.1021","volume":"66","author":[{"given":"Xin","family":"Wang","sequence":"first","affiliation":[{"name":"Beijing University of Chemical Technology , , P.O. Box 53, 15 BeiSanHuan East Road , ,","place":["Beijing, P. R. China, 100029"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hongyan","family":"Yin","sequence":"additional","affiliation":[{"name":"Beijing University of Chemical Technology , , P.O. Box 53, 15 BeiSanHuan East Road , ,","place":["Beijing, P. R. China, 100029"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yekai","family":"Shen","sequence":"additional","affiliation":[{"name":"Beijing University of Chemical Technology , , P.O. Box 53, 15 BeiSanHuan East Road , ,","place":["Beijing, P. R. China, 100029"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4332-4243","authenticated-orcid":true,"given":"Yushan","family":"Zhu","sequence":"additional","affiliation":[{"name":"Beijing University of Chemical Technology , , 15 BeiSanHuan East Road , ,","place":["Beijing, P. R. China, 100029"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6855-0012","authenticated-orcid":true,"given":"Igor V.","family":"Tetko","sequence":"additional","affiliation":[{"name":"Helmholtz Munich - German Research Center for Environmental Health (GmbH) , , Ingolsta\u0308dter Landstra\u00dfe 1 , ,","place":["Neuherberg, Germany, 85764"]},{"name":"BIGCHEM GmbH , Valerystr. 49 , ,","place":["Unterschlei\u00dfheim, Germany, 85716"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4811-8510","authenticated-orcid":true,"given":"Aixia","family":"Yan","sequence":"additional","affiliation":[{"name":"Beijing University of Chemical Technology , , P.O. Box 53, 15 BeiSanHuan East Road , ,","place":["Beijing, P. R. China, 100029"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"316","published-online":{"date-parts":[[2026,9,29]]},"reference":[{"issue":"11","key":"2026101208082204000_cit1","doi-asserted-by":"publisher","first-page":"1211","DOI":"10.16418\/j.issn.1000-3045.2018.11.009","article-title":"Current status and prospects of Industrial\napplications of synthetic biology","volume":"33","author":"Zeng","year":"2018","journal-title":"Bull. Chin.\nAcad. Sci"},{"issue":"8","key":"2026101208082204000_cit2","doi-asserted-by":"publisher","first-page":"1122","DOI":"10.1039\/C9NP00071B","article-title":"Engineering biosynthetic enzymes\nfor industrial natural product synthesis","volume":"37","author":"Galanie","year":"2020","journal-title":"Nat.\nProd. Rep"},{"issue":"7446","key":"2026101208082204000_cit3","doi-asserted-by":"publisher","first-page":"528","DOI":"10.1038\/nature12051","article-title":"High-level semi-synthetic production of the potent\nantimalarial artemisinin","volume":"496","author":"Paddon","year":"2013","journal-title":"Nature"},{"issue":"11","key":"2026101208082204000_cit4","doi-asserted-by":"publisher","first-page":"1914","DOI":"10.3390\/foods14111914","article-title":"Research advances in the synthesis,\nmetabolism, and function of chlorogenic acid","volume":"14","author":"He","year":"2025","journal-title":"Foods"},{"issue":"4","key":"2026101208082204000_cit5","doi-asserted-by":"publisher","first-page":"812","DOI":"10.1039\/d0qo00946f","article-title":"Mapping the resources\nand approaches facilitating computer-aided synthesis planning","volume":"8","author":"Wang","year":"2021","journal-title":"Org. Chem. Front"},{"issue":"11","key":"2026101208082204000_cit6","doi-asserted-by":"publisher","first-page":"3525","DOI":"10.1039\/d0cs00098a","article-title":"QSAR without borders","volume":"49","author":"Muratov","year":"2020","journal-title":"Chem. Soc. Rev"},{"key":"2026101208082204000_cit7","doi-asserted-by":"publisher","first-page":"2476","DOI":"10.3762\/bjoc.20.212","article-title":"Machine learning-guided strategies for reaction conditions design\nand optimization","volume":"20","author":"Chen","year":"2024","journal-title":"Beilstein J. Org. Chem"},{"issue":"1","key":"2026101208082204000_cit8","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1002\/wcms.1694","article-title":"Recent advances\nin deep learning\nfor retrosynthesis","volume":"14","author":"Zhong","year":"2024","journal-title":"WIREs Comput. Mol. Sci"},{"key":"2026101208082204000_cit9","doi-asserted-by":"publisher","first-page":"273","DOI":"10.1016\/j.cjche.2024.10.014","article-title":"Machine learning-assisted\nretrosynthesis planning: Current status and future prospects","volume":"77","author":"Wei","year":"2025","journal-title":"Chin. J. Chem. Eng"},{"issue":"1","key":"2026101208082204000_cit10","doi-asserted-by":"publisher","first-page":"42","DOI":"10.1021\/acs.jcim.3c01524","article-title":"When yield\nprediction does not yield prediction: An overview of the current challenges","volume":"64","author":"Voinarovska","year":"2024","journal-title":"J. Chem. Inf. Model"},{"issue":"3","key":"2026101208082204000_cit11","doi-asserted-by":"publisher","first-page":"558","DOI":"10.1039\/D3DD00252G","article-title":"Models\nMatter: The\nimpact of single-step retrosynthesis on synthesis planning","volume":"3","author":"Torren-Peraire","year":"2024","journal-title":"Digital Discovery"},{"issue":"D1","key":"2026101208082204000_cit12","doi-asserted-by":"publisher","first-page":"D353","DOI":"10.1093\/nar\/gkw1092","article-title":"New perspectives\non genomes, pathways, diseases and drugs","volume":"45","author":"Kanehisa","year":"2017","journal-title":"Nucleic\nAcids Res"},{"issue":"D1","key":"2026101208082204000_cit13","doi-asserted-by":"publisher","first-page":"D693","DOI":"10.1093\/nar\/gkab1016","article-title":"the reaction\nknowledgebase in 2022","volume":"50","author":"Bansal","year":"2022","journal-title":"Nucleic Acids Res"},{"issue":"W1","key":"2026101208082204000_cit14","doi-asserted-by":"publisher","first-page":"W299","DOI":"10.1093\/nar\/gkae405","article-title":"REME: An integrated\nplatform for reaction enzyme mining and evaluation","volume":"52","author":"Shi","year":"2024","journal-title":"Nucleic Acids Res"},{"issue":"10","key":"2026101208082204000_cit15","doi-asserted-by":"publisher","first-page":"501","DOI":"10.1016\/j.tibtech.2010.07.002","article-title":"DREAMS of\nmetabolism","volume":"28","author":"Soh","year":"2010","journal-title":"Trends\nBiotechnol"},{"issue":"11","key":"2026101208082204000_cit16","doi-asserted-by":"publisher","first-page":"1446","DOI":"10.1002\/biot.201400055","article-title":"Validation\nof RetroPath, a computer-aided design tool for metabolic pathway engineering","volume":"9","author":"Fehe\u0301r","year":"2014","journal-title":"Biotechnol. J"},{"key":"2026101208082204000_cit17","doi-asserted-by":"publisher","first-page":"W138","DOI":"10.1093\/nar\/gkq318","article-title":"PathPred: An enzyme-catalyzed\nmetabolic pathway prediction server","volume":"38","author":"Moriya","year":"2010","journal-title":"Nucleic\nAcids Res"},{"key":"2026101208082204000_cit18","doi-asserted-by":"publisher","first-page":"99","DOI":"10.1016\/j.cbpa.2015.06.025","article-title":"Design of\ncomputational retrobiosynthesis tools for\nthe design of de novo synthetic pathways","volume":"28","author":"Hadadi","year":"2015","journal-title":"Curr.\nOpin. Chem. Biol"},{"issue":"D1","key":"2026101208082204000_cit19","doi-asserted-by":"publisher","first-page":"D1229","DOI":"10.1093\/nar\/gky940","article-title":"RetroRules: A database\nof reaction rules for engineering biology","volume":"47","author":"Duigou","year":"2019","journal-title":"Nucleic\nAcids Res"},{"issue":"5","key":"2026101208082204000_cit20","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1002\/biot.202000605","article-title":"A deep\nlearning approach to evaluate the feasibility of enzymatic reactions\ngenerated by retrobiosynthesis","volume":"16","author":"Kim","year":"2021","journal-title":"Biotechnol.\nJ"},{"key":"2026101208082204000_cit21","article-title":"AMTL-RFC: A multi-task learning based method for\nevaluating the feasibility of enzymatic reactions","volume-title":"2023 IEEE International Conference on Bioinformatics and\nBiomedicine (BIBM)","author":"Liu","year":"2023"},{"issue":"2","key":"2026101208082204000_cit22","doi-asserted-by":"publisher","first-page":"129","DOI":"10.1039\/D4ME00118D","article-title":"DORA-XGB: An improved enzymatic reaction\nfeasibility\nclassifier trained using a novel synthetic data approach","volume":"10","author":"Chainani","year":"2025","journal-title":"Mol. Syst. Des. Eng"},{"issue":"1","key":"2026101208082204000_cit23","doi-asserted-by":"publisher","first-page":"4839","DOI":"10.1038\/s41467-025-59827-7","article-title":"Designing\npathways for bioproducing\ncomplex chemicals by combining tools for pathway extraction and ranking","volume":"16","author":"Sveshnikova","year":"2025","journal-title":"Nat. Commun"},{"issue":"4","key":"2026101208082204000_cit24","doi-asserted-by":"publisher","first-page":"243","DOI":"10.1016\/j.synbio.2017.11.002","article-title":"A review\nof computational\ntools for design and reconstruction of metabolic pathways","volume":"2","author":"Wang","year":"2017","journal-title":"Synth. Syst. Biotechnol"},{"issue":"1","key":"2026101208082204000_cit25","doi-asserted-by":"publisher","first-page":"34","DOI":"10.1093\/bioinformatics\/16.1.34","article-title":"History of the enzyme\nnomenclature system","volume":"16","author":"Tipton","year":"2000","journal-title":"Bioinformatics"},{"issue":"1","key":"2026101208082204000_cit26","doi-asserted-by":"publisher","first-page":"2","DOI":"10.1186\/s13321-024-00944-8","article-title":"CLAIRE: A contrastive learning-based\npredictor for EC number of chemical reactions","volume":"17","author":"Zeng","year":"2025","journal-title":"J. Cheminf"},{"issue":"1","key":"2026101208082204000_cit27","doi-asserted-by":"publisher","first-page":"964","DOI":"10.1038\/s41467-022-28536-w","article-title":"Biocatalysed synthesis planning using\ndata-driven learning","volume":"13","author":"Probst","year":"2022","journal-title":"Nat. Commun"},{"key":"2026101208082204000_cit28","doi-asserted-by":"publisher","DOI":"10.26434\/chemrxiv.12298559.v1","article-title":"Unsupervised attention-guided\natom-mapping, Ver. 1","author":"Schwaller","year":"2020","journal-title":"ChemRxiv"},{"key":"2026101208082204000_cit29","unstructured":"Kenneth, R.\n          \n          Requests: Http for humans; https:\/\/docs.python-requests.org, Accessed 10 October 2024."},{"issue":"8090","key":"2026101208082204000_cit30","doi-asserted-by":"publisher","first-page":"639","DOI":"10.1038\/s41586-025-09697-2","article-title":"Enzyme specificity prediction using\ncross-attention graph neural networks","volume":"647","author":"Cui","year":"2025","journal-title":"Nature"},{"key":"2026101208082204000_cit31","unstructured":"Greg, L.\n          \n          RDKit: Open-source cheminformatics\nsoftware; http:\/\/www.rdkit.org Accessed 1 May 2025."},{"issue":"1","key":"2026101208082204000_cit32","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s13321-020-0416-x","article-title":"Visualization of\nvery large high-dimensional data sets\nas minimum spanning trees","volume":"12","author":"Probst","year":"2020","journal-title":"J. Cheminf."},{"issue":"11","key":"2026101208082204000_cit33","doi-asserted-by":"publisher","first-page":"103356","DOI":"10.1016\/j.drudis.2022.103356","article-title":"Concepts and applications\nof chemical fingerprint for hit and lead screening","volume":"27","author":"Yang","year":"2022","journal-title":"Drug Discovery Today"},{"issue":"2","key":"2026101208082204000_cit34","doi-asserted-by":"publisher","first-page":"693","DOI":"10.1021\/acs.jcim.4c02097","article-title":"Active learning to select the most\nsuitable reagents and one-step organic chemistry reactions for prioritizing\ntarget-specific hits from ultralarge chemical spaces","volume":"65","author":"Kozyrev","year":"2025","journal-title":"J. Chem. Inf. Model"},{"issue":"5","key":"2026101208082204000_cit35","doi-asserted-by":"publisher","first-page":"742","DOI":"10.1021\/ci100050t","article-title":"Extended-connectivity\nfingerprints","volume":"50","author":"Rogers","year":"2010","journal-title":"J. Chem.\nInf. Model"},{"issue":"2","key":"2026101208082204000_cit36","doi-asserted-by":"publisher","first-page":"144","DOI":"10.1038\/s42256-020-00284-w","article-title":"Mapping the space of chemical reactions\nusing attention-based neural networks","volume":"3","author":"Schwaller","year":"2021","journal-title":"Nat.\nMach. Intell"},{"issue":"2","key":"2026101208082204000_cit37","doi-asserted-by":"publisher","first-page":"91","DOI":"10.1039\/D1DD00006C","article-title":"Reaction classification\nand yield\nprediction using the differential reaction fingerprint DRFP","volume":"1","author":"Probst","year":"2022","journal-title":"Digital Discovery"},{"issue":"12","key":"2026101208082204000_cit38","doi-asserted-by":"publisher","first-page":"2353","DOI":"10.1021\/acs.jcim.6b00591","article-title":"Extreme\ngradient boosting as a method for quantitative structure-activity\nrelationships","volume":"56","author":"Sheridan","year":"2016","journal-title":"J. Chem. Inf. Model"},{"key":"2026101208082204000_cit39","unstructured":"Tianqi, C.\n          \n          XGBoost: A scalable tree\nboosting system; https:\/\/xgboost.ai Accessed 10 May 2025."},{"key":"2026101208082204000_cit40","unstructured":"Optuna: An end-to-end hyperparameter\noptimization framework; Preferred Networks,\nInc.: https:\/\/optuna.org Accessed 11 May\n2025."},{"issue":"5","key":"2026101208082204000_cit41","doi-asserted-by":"publisher","first-page":"826","DOI":"10.1021\/ci00027a006","article-title":"Neural network studies. 1. Comparison\nof overfitting and overtraining","volume":"35","author":"Tetko","year":"1995","journal-title":"J. Chem. Inf.\nComput. Sci"},{"key":"2026101208082204000_cit42","doi-asserted-by":"publisher","first-page":"11","DOI":"10.1016\/j.neucom.2016.12.038","article-title":"A survey of deep neural network architectures and their\napplications","volume":"234","author":"Liu","year":"2017","journal-title":"Neurocomputing"},{"key":"2026101208082204000_cit43","article-title":"PyTorch: An\nimperative style, high-performance\ndeep learning library","volume-title":"Proceedings of the\n33rd International Conference on Neural Information Processing Systems\n(NeurIPS 2019)","author":"Paszke","year":"2019"},{"key":"2026101208082204000_cit44","article-title":"Attention is all you need","volume-title":"Proceedings of the 31st International Conference on\nNeural Information Processing Systems (NeurIPS 2017)","author":"Vaswani","year":"2017"},{"key":"2026101208082204000_cit45","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2209.01712","article-title":"ChemBERTa-2: Towards chemical foundation\nmodels, Ver. 1","author":"Ahmad","year":"2022","journal-title":"arXiv"},{"key":"2026101208082204000_cit46","article-title":"Transformers: State-of-the-art natural language processing","volume-title":"Proceedings of the 2020 Conference on Empirical Methods\nin Natural Language Processing (EMNLP)","author":"Wolf","year":"2020"},{"issue":"1","key":"2026101208082204000_cit47","doi-asserted-by":"publisher","first-page":"73","DOI":"10.1002\/minf.201300030","article-title":"Modeling the biodegradability of chemical compounds\nusing the online chemical modeling environment (OCHEM)","volume":"33","author":"Vorberg","year":"2014","journal-title":"Mol. Inform"},{"issue":"3","key":"2026101208082204000_cit48","doi-asserted-by":"publisher","first-page":"837","DOI":"10.2307\/2531595","article-title":"Comparing the areas under two or\nmore correlated receiver operating characteristic curves: A nonparametric\napproach","volume":"44","author":"DeLong","year":"1988","journal-title":"Biometrics"},{"issue":"8","key":"2026101208082204000_cit49","doi-asserted-by":"publisher","first-page":"775","DOI":"10.1016\/S0006-3223(98)00043-2","article-title":"Multiple correlations\nand bonferroni\u2019s correction","volume":"44","author":"Curtin","year":"1998","journal-title":"Biol. Psychiatry"},{"issue":"12","key":"2026101208082204000_cit50","doi-asserted-by":"publisher","first-page":"3552","DOI":"10.1039\/D5DD00369E","article-title":"MolEncoder:\ntowards optimal masked language modeling for molecules","volume":"4","author":"Kru\u0308ger","year":"2025","journal-title":"Digit. Discov."},{"issue":"1","key":"2026101208082204000_cit51","doi-asserted-by":"publisher","first-page":"1931","DOI":"10.1038\/s41467-023-37558-x","article-title":"Construction of a synthetic\nmetabolic pathway for biosynthesis of 2,4-dihydroxybutyric acid from\nethylene glycol","volume":"14","author":"Fraza\u0303o","year":"2023","journal-title":"Nat. Commun"}],"container-title":["Journal of Chemical Information\nand Modeling"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/pubs.acs.org\/doi\/pdf\/10.1021\/acs.jcim.6c02294","content-type":"application\/pdf","content-version":"vor","intended-application":"unspecified"},{"URL":"https:\/\/pubs.acs.org\/jcisd8\/article-pdf\/66\/19\/12613\/68685360\/acs.jcim.6c02294.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/pubs.acs.org\/jcisd8\/article-pdf\/66\/19\/12613\/68685360\/acs.jcim.6c02294.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,12]],"date-time":"2026-10-12T12:30:21Z","timestamp":1791808221000},"score":1,"resource":{"primary":{"URL":"https:\/\/pubs.acs.org\/jcisd8\/article\/66\/19\/12613\/5440449\/Enzymatic-Reaction-Feasibility-Classification"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,29]]},"references-count":51,"journal-issue":{"issue":"19","published-online":{"date-parts":[[2026,9,29]]},"published-print":{"date-parts":[[2026,10,12]]}},"URL":"https:\/\/doi.org\/10.1021\/acs.jcim.6c02294","relation":{},"ISSN":["1549-9596","1549-960X"],"issn-type":[{"value":"1549-9596","type":"print"},{"value":"1549-960X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,29]]}}}