{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,23]],"date-time":"2026-10-23T19:15:58Z","timestamp":1792782958347,"version":"4.5.0"},"reference-count":30,"publisher":"Springer Science and Business Media LLC","issue":"7","license":[{"start":{"date-parts":[[2026,9,1]],"date-time":"2026-09-01T00:00:00Z","timestamp":1788220800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,9,1]],"date-time":"2026-09-01T00:00:00Z","timestamp":1788220800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"name":"Universidade Federal Do Vale Do S\u00e3o Francisco","id":[{"id":"https:\/\/ror.org\/00devjr72","id-type":"ROR","asserted-by":"crossref"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Ecotoxicology"],"published-print":{"date-parts":[[2026,9]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    Honey bee (\n                    <jats:italic>Apis mellifera<\/jats:italic>\n                    ) toxicity assessment is essential for environmental risk evaluation of agrochemicals, yet experimental data are costly and often limited, constraining the development of robust predictive models. This study investigates whether transfer learning via pretrained chemical language models can provide competitive QSAR performance for bee toxicity classification. Using the ApisTox dataset (1,035 compounds; 296 toxic and 739 non-toxic for bees), molecular representations were derived from SMILES in three ways: (i) PaDEL molecular descriptors, (ii) RDKit Morgan fingerprints, and (iii) MolFormer embeddings extracted from a publicly available reduced-scale pretrained checkpoint (~\u2009100\u00a0M molecules; ~10% ZINC\u2009+\u2009~\u200910% PubChem), used as frozen features. Three classical classifiers: Random Forest, Support Vector Machine, and Multilayer Perceptron were trained and evaluated under 5-fold cross-validation. Fingerprints paired with RF achieved the best overall discrimination, as measured by the Area Under the Receiver Operating Characteristic (ROC-AUC\u2009=\u20090.866). Importantly, MolFormer embeddings combined with SVM reached near-parity (ROC-AUC\u2009=\u20090.859) and consistently outperformed PaDEL descriptors for all classifiers (\u0394ROC-AUC\u2009=\u2009+\u20090.005 to +\u20090.021). These results demonstrate that transfer-learned chemical embeddings can rival established QSAR baselines while simplifying feature engineering, supporting their practical adoption for ecotoxicological screening under limited labeled data.\n                  <\/jats:p>","DOI":"10.1007\/s10646-026-03149-x","type":"journal-article","created":{"date-parts":[[2026,9,1]],"date-time":"2026-09-01T16:32:56Z","timestamp":1788280376000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Transfer learning for honey bee toxicity prediction: MolFormer versus classical QSAR representations"],"prefix":"10.1007","volume":"35","author":[{"given":"Alan Victor","family":"de Souza Pinho","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Edilson Beserra","family":"de Alencar Filho","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Rosalvo Ferreira","family":"de Oliveira Neto","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,9,1]]},"reference":[{"key":"3149_CR2","doi-asserted-by":"publisher","DOI":"10.5281\/zenodo.11062076","author":"J Adamczyk","year":"2024","unstructured":"Adamczyk J, Poziemski J, Siedlecki P (2024) ApisTox dataset. Zenodo. https:\/\/doi.org\/10.5281\/zenodo.11062076","journal-title":"Zenodo"},{"key":"3149_CR1","doi-asserted-by":"publisher","first-page":"5","DOI":"10.1038\/s41597-024-04232-w","volume":"12","author":"J Adamczyk","year":"2025","unstructured":"Adamczyk J, Poziemski J, Siedlecki P (2025) ApisTox: a new benchmark dataset for the classification of small molecules toxicity on honey bees. Sci Data 12:5. https:\/\/doi.org\/10.1038\/s41597-024-04232-w","journal-title":"Sci Data"},{"issue":"16","key":"3149_CR3","doi-asserted-by":"publisher","first-page":"8683","DOI":"10.1021\/acs.jmedchem.9b02147","volume":"63","author":"C Cai","year":"2020","unstructured":"Cai C, Wang S, Xu Y, Zhang W, Tang K, Ouyang Q, Lai L, Pei J (2020) Transfer learning for drug discovery. J Med Chem 63(16):8683\u20138694. https:\/\/doi.org\/10.1021\/acs.jmedchem.9b02147","journal-title":"J Med Chem"},{"key":"3149_CR4","doi-asserted-by":"publisher","first-page":"4977","DOI":"10.1021\/jm4004285","volume":"57","author":"A Cherkasov","year":"2014","unstructured":"Cherkasov A, Muratov EN, Fourches D et al (2014) QSAR modeling: where have you been? Where are you going to? J Med Chem 57:4977\u20135010. https:\/\/doi.org\/10.1021\/jm4004285","journal-title":"J Med Chem"},{"key":"3149_CR5","doi-asserted-by":"publisher","first-page":"2028","DOI":"10.1007\/s10646-025-02970-0","volume":"34","author":"TC Dami\u00e3o","year":"2025","unstructured":"Dami\u00e3o TC, Oliveira Neto RF, Alencar Filho EB (2025) New generation of QSAR modeling for bee safety: predicting toxicity using graph neural networks and ApisTox data. Ecotoxicology 34:2028\u20132039. https:\/\/doi.org\/10.1007\/s10646-025-02970-0","journal-title":"Ecotoxicology"},{"issue":"7","key":"3149_CR6","doi-asserted-by":"publisher","first-page":"3295","DOI":"10.2903\/j.efsa.2013.3295","volume":"11","author":"EFSA","year":"2013","unstructured":"EFSA (2013) Guidance on the risk assessment of plant protection products on bees (Apis mellifera, Bombus spp. and solitary bees). EFSA J 11(7):3295. https:\/\/doi.org\/10.2903\/j.efsa.2013.3295","journal-title":"EFSA J"},{"key":"3149_CR7","doi-asserted-by":"publisher","first-page":"861","DOI":"10.1016\/j.patrec.2005.10.010","volume":"27","author":"T Fawcett","year":"2006","unstructured":"Fawcett T (2006) An introduction to ROC analysis. Pattern Recognit Lett 27:861\u2013874. https:\/\/doi.org\/10.1016\/j.patrec.2005.10.010","journal-title":"Pattern Recognit Lett"},{"key":"3149_CR8","doi-asserted-by":"publisher","first-page":"1283","DOI":"10.1007\/s11030-021-10250-2","volume":"25","author":"C Hung","year":"2021","unstructured":"Hung C, Gini G (2021) QSAR modeling without descriptors using graph convolutional neural networks: the case of mutagenicity prediction. Mol Divers 25:1283\u20131299. https:\/\/doi.org\/10.1007\/s11030-021-10250-2","journal-title":"Mol Divers"},{"key":"3149_CR9","first-page":"1137","volume":"2","author":"R Kohavi","year":"1995","unstructured":"Kohavi R (1995) A study of cross-validation and bootstrap for accuracy estimation and model selection. Proc IJCAI 2:1137\u20131143","journal-title":"Proc IJCAI"},{"key":"3149_CR10","doi-asserted-by":"publisher","DOI":"10.5281\/zenodo.15286010","author":"G Landrum","year":"2025","unstructured":"Landrum G et al (2025) RDKit: Open-source cheminformatics software. Zenodo. https:\/\/doi.org\/10.5281\/zenodo.15286010","journal-title":"Zenodo"},{"issue":"4","key":"3149_CR11","doi-asserted-by":"publisher","first-page":"1050","DOI":"10.1080\/10807039.2015.1133242","volume":"22","author":"KA Lewis","year":"2016","unstructured":"Lewis KA, Tzilivakis J, Warner DJ, Green A (2016) An international database for pesticide risk assessments and management. Hum Ecol Risk Assess Int J 22(4):1050\u20131064. https:\/\/doi.org\/10.1080\/10807039.2015.1133242","journal-title":"Hum Ecol Risk Assess Int J"},{"key":"3149_CR12","doi-asserted-by":"publisher","first-page":"27","DOI":"10.1186\/s13321-020-00430-x","volume":"12","author":"X Li","year":"2020","unstructured":"Li X, Fourches D (2020) Inductive transfer learning for molecular activity prediction: next-gen QSAR models with MolPMoFiT. J Cheminform 12:27. https:\/\/doi.org\/10.1186\/s13321-020-00430-x","journal-title":"J Cheminform"},{"key":"3149_CR13","doi-asserted-by":"publisher","DOI":"10.1787\/9789264070165-en","volume-title":"Test No. 213: Honeybees, Acute Oral Toxicity Test, OECD Guidelines for the Testing of Chemicals, Section 2","author":"OECD","year":"1998","unstructured":"OECD (1998a) Test No. 213: Honeybees, Acute Oral Toxicity Test, OECD Guidelines for the Testing of Chemicals, Sect.\u00a0 2. OECD Publishing, Paris. https:\/\/doi.org\/10.1787\/9789264070165-en"},{"key":"3149_CR14","doi-asserted-by":"publisher","DOI":"10.1787\/9789264070189-en","volume-title":"Test No. 214: Honeybees, Acute Contact Toxicity Test, OECD Guidelines for the Testing of Chemicals, Section 2","author":"OECD","year":"1998","unstructured":"OECD (1998b) Test No. 214: Honeybees, Acute Contact Toxicity Test, OECD Guidelines for the Testing of Chemicals, Sect.\u00a0 2. OECD Publishing, Paris. https:\/\/doi.org\/10.1787\/9789264070189-en"},{"key":"3149_CR15","doi-asserted-by":"publisher","DOI":"10.1787\/9789264085442-en","volume-title":"Guidance document on the validation of (quantitative) structure-activity relationship [(Q)SAR] models","author":"OECD","year":"2014","unstructured":"OECD (2014) Guidance document on the validation of (quantitative) structure-activity relationship [(Q)SAR] models. OECD, Paris"},{"key":"3149_CR16","doi-asserted-by":"publisher","first-page":"1520","DOI":"10.1002\/etc.5324","volume":"41","author":"JH Olker","year":"2022","unstructured":"Olker JH, Valenti TW, Elonen CM et al (2022) The ECOTOXicology Knowledgebase: A Curated Database of Ecologically Relevant Toxicity Tests to Support Environmental Research and Risk Assessment. Environ Toxicol Chem 41:1520\u20131537. https:\/\/doi.org\/10.1002\/etc.5324","journal-title":"Environ Toxicol Chem"},{"key":"3149_CR17","doi-asserted-by":"publisher","first-page":"1345","DOI":"10.1109\/TKDE.2009.191","volume":"22","author":"SJ Pan","year":"2010","unstructured":"Pan SJ, Yang Q (2010) A survey on transfer learning. IEEE Trans Knowl Data Eng 22:1345\u20131359. https:\/\/doi.org\/10.1109\/TKDE.2009.191","journal-title":"IEEE Trans Knowl Data Eng"},{"key":"3149_CR18","first-page":"2825","volume":"12","author":"F Pedregosa","year":"2011","unstructured":"Pedregosa F et al (2011) Scikit-learn: machine learning in Python. J Mach Learn Res 12:2825\u20132830","journal-title":"J Mach Learn Res"},{"key":"3149_CR19","doi-asserted-by":"publisher","first-page":"345","DOI":"10.1016\/j.tree.2010.01.007","volume":"25","author":"SG Potts","year":"2010","unstructured":"Potts SG, Biesmeijer JC, Kremen C et al (2010) Global pollinator declines: trends, impacts and drivers. Trends Ecol Evol 25:345\u2013353. https:\/\/doi.org\/10.1016\/j.tree.2010.01.007","journal-title":"Trends Ecol Evol"},{"key":"3149_CR20","volume-title":"Recent advances in QSAR studies: methods and applications","year":"2010","unstructured":"Puzyn T, Leszczynski J, Cronin MTD (eds) (2010) Recent advances in QSAR studies: methods and applications. Springer, Dordrecht"},{"key":"3149_CR21","doi-asserted-by":"publisher","first-page":"2514","DOI":"10.1039\/D4SC03921A","volume":"16","author":"MC Ramos","year":"2025","unstructured":"Ramos MC, Collison CJ, White AD (2025) A review of large language models and autonomous agents in chemistry. Chem Sci 16:2514\u20132572. https:\/\/doi.org\/10.1039\/D4SC03921A","journal-title":"Chem Sci"},{"issue":"12","key":"3149_CR22","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(12):1256\u20131264. https:\/\/doi.org\/10.1038\/s42256-022-00580-7","journal-title":"Nat Mach Intell"},{"key":"3149_CR23","doi-asserted-by":"publisher","first-page":"120","DOI":"10.1021\/acscentsci.7b00512","volume":"4","author":"MHS Segler","year":"2018","unstructured":"Segler MHS, Kogej T, Tyrchan C, Waller MP (2018) Generating focused molecule libraries for drug discovery with recurrent neural networks. ACS Cent Sci 4:120\u2013131. https:\/\/doi.org\/10.1021\/acscentsci.7b00512","journal-title":"ACS Cent Sci"},{"key":"3149_CR24","doi-asserted-by":"publisher","first-page":"74","DOI":"10.3389\/fphar.2018.00074","volume":"9","author":"RS Sim\u00f5es","year":"2018","unstructured":"Sim\u00f5es RS, Maltarollo VG, Oliveira PR, Honorio KM (2018) Transfer and multi-task learning in QSAR modeling: advances and challenges. Front Pharmacol 9:74. https:\/\/doi.org\/10.3389\/fphar.2018.00074","journal-title":"Front Pharmacol"},{"key":"3149_CR25","doi-asserted-by":"publisher","first-page":"476","DOI":"10.1002\/minf.201000061","volume":"29","author":"A Tropsha","year":"2010","unstructured":"Tropsha A (2010) Best practices for QSAR model development, validation, and exploitation. Mol Inf 29:476\u2013488. https:\/\/doi.org\/10.1002\/minf.201000061","journal-title":"Mol Inf"},{"key":"3149_CR28","unstructured":"U.S. EPA (2014) Guidance for assessing pesticide risks to bees. Office of Pesticide Programs, United States Environmental Protection Agency, Washington, DC. Available at: https:\/\/www.epa.gov\/sites\/default\/files\/2014-06\/documents\/pollinator_risk_assessment_guidance_06_19_14.pdf"},{"key":"3149_CR26","doi-asserted-by":"publisher","first-page":"85","DOI":"10.1038\/s42004-024-01169-4","volume":"7","author":"S Yao","year":"2024","unstructured":"Yao S, Song J, Jia L et al (2024) Fast and effective molecular property prediction with transferability map. Commun Chem 7:85. https:\/\/doi.org\/10.1038\/s42004-024-01169-4","journal-title":"Commun Chem"},{"key":"3149_CR27","doi-asserted-by":"publisher","first-page":"1466","DOI":"10.1002\/jcc.21707","volume":"32","author":"CW Yap","year":"2011","unstructured":"Yap CW (2011) PaDEL-descriptor: an open source software to calculate molecular descriptors and fingerprints. J Comput Chem 32:1466\u20131474. https:\/\/doi.org\/10.1002\/jcc.21707","journal-title":"J Comput Chem"},{"key":"3149_CR29","doi-asserted-by":"publisher","first-page":"89","DOI":"10.1186\/s13321-024-00886-1","volume":"16","author":"C Zhang","year":"2024","unstructured":"Zhang C, Zhai Y, Gong Z et al (2024) Transfer learning across different chemical domains: virtual screening of organic materials with deep learning models pretrained on small molecule and chemical reaction data. J Cheminform 16:89. https:\/\/doi.org\/10.1186\/s13321-024-00886-1","journal-title":"J Cheminform"},{"key":"3149_CR30","doi-asserted-by":"publisher","first-page":"1191","DOI":"10.1007\/s11427-018-9342-2","volume":"61","author":"F Zhong","year":"2018","unstructured":"Zhong F, Xing J, Li X et al (2018) Artificial intelligence in drug design. Sci China Life Sci 61:1191\u20131204. https:\/\/doi.org\/10.1007\/s11427-018-9342-27.Statements & Declarations","journal-title":"Sci China Life Sci"}],"container-title":["Ecotoxicology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10646-026-03149-x.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s10646-026-03149-x","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10646-026-03149-x.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,26]],"date-time":"2026-09-26T17:02:23Z","timestamp":1790442143000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s10646-026-03149-x"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9]]},"references-count":30,"journal-issue":{"issue":"7","published-print":{"date-parts":[[2026,9]]}},"alternative-id":["3149"],"URL":"https:\/\/doi.org\/10.1007\/s10646-026-03149-x","relation":{},"ISSN":["0963-9292","1573-3017"],"issn-type":[{"value":"0963-9292","type":"print"},{"value":"1573-3017","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9]]},"assertion":[{"value":"17 March 2026","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"5 August 2026","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"1 September 2026","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"During the preparation of this work the author(s) used Chat GPT 5.2 (Open AI) in order to evaluate and inspire some texts and codes. After using this tool\/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.","order":1,"name":"Ethics","label":"Declaration of generative AI and AI-assisted technologies in the writing process","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"This paper does not contain any studies with human participants or animals performed by any of the authors.","order":2,"name":"Ethics","label":"Ethics approval","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors declare no competing interests.","order":3,"name":"Ethics","label":"Conflict of interests","group":{"name":"EthicsHeading","label":"Declarations"}}],"article-number":"166"}}