{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,14]],"date-time":"2026-09-14T14:12:00Z","timestamp":1789395120366,"version":"build-2803163510"},"reference-count":44,"publisher":"American Chemical Society (ACS)","issue":"17","license":[{"start":{"date-parts":[[2026,9,1]],"date-time":"2026-09-01T00:00:00Z","timestamp":1788220800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,9,1]],"date-time":"2026-09-01T00:00:00Z","timestamp":1788220800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"},{"start":{"date-parts":[[2026,9,1]],"date-time":"2026-09-01T00:00:00Z","timestamp":1788220800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-045"}],"funder":[{"DOI":"10.13039\/na","name":"Young Elite Scientists Sponsorship Program by CAST","doi-asserted-by":"publisher","award":["Grant No. 2024QNRC001"],"award-info":[{"award-number":["Grant No. 2024QNRC001"]}],"id":[{"id":"10.13039\/na","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,9,14]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Molecular representations that capture both chemical information and property-relevant differences are essential for reliable molecular property prediction in drug discovery and toxicity assessments. Fragment-level SMILES representation learning is effective at modeling substructure semantics, but structural reconstruction objectives alone do not explicitly encode the descriptor-derived physicochemical information needed for downstream prediction, especially in small data and scaffold-split settings. To address this limitation, we propose PKP-DiffMol, a physicochemical knowledge-prompt encoding and latent diffusion framework built upon the pretrained SMI-EDITOR encoder. PKP-DiffMol improves fragment-level molecular representations by using numerical and semantic physicochemical knowledge. The numerical physicochemical prior branch encodes RDKit descriptors as continuous quantitative priors, while the physicochemical knowledge prompt encoder transforms the same descriptors into semantic physicochemical representations by using a Transformer-based text encoder. The two complementary views are then integrated through hierarchical physicochemical knowledge fusion to construct a chemically enriched, fused latent space. In this space, a quality-controlled conditional latent diffusion module learns label-conditioned distributions, generates synthetic molecular latent representations, and retains reliable samples through a Mahalanobis distance quality gate. Experiments on seven MoleculeNet classification datasets under scaffold splitting show that PKP-DiffMol increases the mean ROC-AUC from 77.80% for the SMI-EDITOR backbone to 80.53% and obtains the best result on five of the seven datasets. The largest gains are observed on BACE (+6.91%), MUV (+4.29%), and SIDER (+3.92%), showing strong predictive performance across bioactivity prediction, virtual screening, and adverse drug reaction prediction tasks. Further analyses indicate that numerical and semantic physicochemical knowledge improves class separability and the correspondence between representation distances and RDKit descriptor distances, while Mahalanobis-filtered label-conditioned latent augmentation supports prediction-head refinement.<\/jats:p>","DOI":"10.1021\/acs.jcim.6c01547","type":"journal-article","created":{"date-parts":[[2026,9,3]],"date-time":"2026-09-03T10:18:18Z","timestamp":1788430698000},"page":"10737-10753","source":"Crossref","is-referenced-by-count":0,"title":["PKP-Diffmol: A\nPhysicochemical Knowledge-Prompt Encoding\nand Latent Diffusion Framework for Molecular Property Prediction"],"prefix":"10.1021","volume":"66","author":[{"ORCID":"https:\/\/orcid.org\/0009-0009-7148-8959","authenticated-orcid":true,"given":"Ruizi","family":"Liu","sequence":"first","affiliation":[{"name":"Northeast Agricultural University , , ,","place":["Harbin, China, 150030"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8224-9891","authenticated-orcid":true,"given":"Tongtong","family":"Yuan","sequence":"additional","affiliation":[{"name":"Beijing University of Technology , , ,","place":["Beijing, China, 100124"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Molin","family":"Guo","sequence":"additional","affiliation":[{"name":"Kingfa Sci.& Tech. Co., Ltd. , ,","place":["Guangdong, China, 510663"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"316","published-online":{"date-parts":[[2026,9,1]]},"reference":[{"key":"2026091409074041700_cit1","doi-asserted-by":"publisher","first-page":"31","DOI":"10.1021\/ci00057a005","article-title":"SMILES, a\nchemical language and information system. 1. Introduction to methodology\nand encoding rules","volume":"28","author":"Weininger","year":"1988","journal-title":"J. Chem. Inf. Comput. Sci."},{"key":"2026091409074041700_cit2","doi-asserted-by":"publisher","first-page":"513","DOI":"10.1039\/C7SC02664A","article-title":"MoleculeNet: A benchmark\nfor molecular machine learning","volume":"9","author":"Wu","year":"2018","journal-title":"Chem. Sci."},{"key":"2026091409074041700_cit3","doi-asserted-by":"publisher","first-page":"3370","DOI":"10.1021\/acs.jcim.9b00237","article-title":"Analyzing\nLearned Molecular Representations for Property\nPrediction","volume":"59","author":"Yang","year":"2019","journal-title":"J. Chem. Inf. Model."},{"key":"2026091409074041700_cit4","doi-asserted-by":"publisher","first-page":"8749","DOI":"10.1021\/acs.jmedchem.9b00959","article-title":"Pushing the Boundaries\nof Molecular Representation for Drug Discovery\nwith the Graph Attention Mechanism","volume":"63","author":"Xiong","year":"2020","journal-title":"J. Med.\nChem."},{"key":"2026091409074041700_cit5","article-title":"Self-Supervised\nGraph Transformer on Large-Scale Molecular Data","author":"Rong","year":"2020","journal-title":"arXiv"},{"key":"2026091409074041700_cit6","doi-asserted-by":"publisher","first-page":"2579","DOI":"10.1093\/bioinformatics\/btac112","article-title":"Advanced graph\nand sequence neural networks for molecular property prediction and\ndrug discovery","volume":"38","author":"Wang","year":"2022","journal-title":"Bioinformatics"},{"key":"2026091409074041700_cit7","doi-asserted-by":"publisher","first-page":"6259","DOI":"10.1021\/acs.jcim.4c00747","article-title":"Transformers for Molecular\nProperty Prediction: Lessons Learned from the Past Five Years","volume":"64","author":"Sultan","year":"2024","journal-title":"J. Chem. Inf. Model."},{"key":"2026091409074041700_cit8","article-title":"SMI-Editor: Edit-based\nSMILES language model with fragment-level supervision","author":"Zheng","year":"2025","journal-title":"arXiv"},{"key":"2026091409074041700_cit9","doi-asserted-by":"publisher","first-page":"742","DOI":"10.1021\/ci100050t","article-title":"Extended-connectivity fingerprints","volume":"50","author":"Rogers","year":"2010","journal-title":"J. Chem.\nInf. Model."},{"key":"2026091409074041700_cit10","doi-asserted-by":"publisher","first-page":"btaf466","DOI":"10.1093\/bioinformatics\/btaf466","article-title":"MolPrompt: Improving Multi-modal\nMolecular Pre-training with Knowledge Prompts","volume":"41","author":"Li","year":"2025","journal-title":"Bioinformatics"},{"key":"2026091409074041700_cit11","doi-asserted-by":"publisher","first-page":"9173","DOI":"10.1021\/acs.jcim.4c01240","article-title":"From SMILES to Enhanced\nMolecular Property Prediction: A Unified Multimodal Framework with\nPredicted 3D Conformers and Contrastive Learning Techniques","volume":"64","author":"Nguyen","year":"2024","journal-title":"J. Chem. Inf. Model."},{"key":"2026091409074041700_cit12","doi-asserted-by":"publisher","first-page":"5438","DOI":"10.1021\/acs.jcim.5c00577","article-title":"Effective and Explainable\nMolecular\nProperty Prediction by Chain-of-Thought Enabled Large Language Models\nand Multi-Modal Molecular Information Fusion","volume":"65","author":"Jin","year":"2025","journal-title":"J. Chem. Inf. Model."},{"key":"2026091409074041700_cit13","doi-asserted-by":"publisher","first-page":"5163","DOI":"10.1021\/acs.jcim.5c00436","article-title":"Triview Molecular Representation\nLearning Combined with Multitask Optimization for Enhanced Molecular\nProperty Prediction","volume":"65","author":"Han","year":"2025","journal-title":"J. Chem. Inf. Model"},{"key":"2026091409074041700_cit14","doi-asserted-by":"publisher","first-page":"279","DOI":"10.1038\/s42256-022-00447-x","article-title":"Molecular\ncontrastive learning of representations via graph neural networks","volume":"4","author":"Wang","year":"2022","journal-title":"Nat. Mach. Intell."},{"key":"2026091409074041700_cit15","article-title":"Pre-training via\nDenoising\nfor Molecular Property Prediction","author":"Zaidi","year":"2022","journal-title":"arXiv"},{"key":"2026091409074041700_cit16","article-title":"Mole-BERT:\nRethinking pre-training graph neural networks for molecules","author":"Xia","year":"2023","journal-title":"ChemRxiv"},{"key":"2026091409074041700_cit17","article-title":"InfoGraph: Unsupervised and semi-supervised\ngraph-level representation learning via mutual information maximization","author":"Sun","year":"2020","journal-title":"arXiv"},{"key":"2026091409074041700_cit18","article-title":"N-Gram Graph:\nSimple Unsupervised Representation for Graphs, with Applications to\nMolecules","author":"Liu","year":"2019","journal-title":"arXiv"},{"key":"2026091409074041700_cit19","article-title":"Pre-training molecular graph representation\nwith\n3D geometry","author":"Liu","year":"2022","journal-title":"arXiv"},{"key":"2026091409074041700_cit20","article-title":"3D\nInfomax Improves\nGNNs for Molecular Property Prediction","author":"Sta\u0308rk","year":"2022","journal-title":"arXiv"},{"key":"2026091409074041700_cit21","article-title":"A Group Symmetric Stochastic Differential Equation\nModel for Molecule Multi-modal Pretraining","author":"Liu","year":"2023","journal-title":"arXiv"},{"key":"2026091409074041700_cit22","doi-asserted-by":"crossref","first-page":"429","DOI":"10.1145\/3307339.3342186","article-title":"SMILES-BERT: Large scale unsupervised pre-training\nfor molecular property prediction","volume-title":"Proceedings\nof the 10th ACM International Conference on Bioinformatics, Computational\nBiology and Health Informatics","author":"Wang","year":"2019"},{"key":"2026091409074041700_cit23","article-title":"MolTRES: Improving chemical\nlanguage representation\nlearning for molecular property prediction","author":"Park","year":"2024","journal-title":"arXiv"},{"key":"2026091409074041700_cit24","article-title":"Pre-trained molecular language models\nwith random functional group masking","author":"Peng","year":"2024","journal-title":"arXiv"},{"key":"2026091409074041700_cit25","unstructured":"Landrum, G.\n          \n          RDKit: Open-source cheminformatics. https:\/\/www.rdkit.org, 2013; https:\/\/www.rdkit.org, Accessed2026\u201305\u201313."},{"key":"2026091409074041700_cit26","doi-asserted-by":"publisher","first-page":"7568","DOI":"10.1038\/s41467-023-43214-1","article-title":"A knowledge-guided\npre-training framework for improving molecular\nrepresentation learning","volume":"14","author":"Li","year":"2023","journal-title":"Nat. Commun."},{"key":"2026091409074041700_cit27","doi-asserted-by":"publisher","first-page":"9443","DOI":"10.1021\/acs.jcim.5c01068","article-title":"KGG: Knowledge-Guided\nGraph Self-Supervised Learning to Enhance Molecular Property Predictions","volume":"65","author":"To","year":"2025","journal-title":"J. Chem. Inf. Model."},{"key":"2026091409074041700_cit28","doi-asserted-by":"publisher","first-page":"1112","DOI":"10.1021\/acs.jcim.3c01468","article-title":"Molecular Contrastive Pretraining with Collaborative Featurizations","volume":"64","author":"Zhu","year":"2024","journal-title":"J. Chem. Inf. Model."},{"key":"2026091409074041700_cit29","article-title":"Deep Learning Foundation\nModels from Classical Molecular Descriptors","author":"Burns","year":"2025","journal-title":"arXiv"},{"key":"2026091409074041700_cit30","article-title":"Beyond chemical\nlanguage: A multimodal approach to enhance molecular property prediction","author":"Soares","year":"2023","journal-title":"arXiv"},{"key":"2026091409074041700_cit31","doi-asserted-by":"publisher","first-page":"10221","DOI":"10.1021\/acs.jcim.4c02261","article-title":"Improving the Reliability of Molecular String Representations for\nGenerative Chemistry","volume":"65","author":"Reboul","year":"2025","journal-title":"J. Chem. Inf. Model."},{"key":"2026091409074041700_cit32","doi-asserted-by":"publisher","first-page":"5135","DOI":"10.1021\/acs.jcim.6c00314","article-title":"Fragment-Aware Contrastive\nLearning Framework for Molecular Property Prediction","volume":"66","author":"Ru","year":"2026","journal-title":"J. Chem. Inf. Model."},{"key":"2026091409074041700_cit33","doi-asserted-by":"publisher","first-page":"7238","DOI":"10.1021\/acs.jcim.4c01107","article-title":"Diffusion Models in De Novo Drug\nDesign","volume":"64","author":"Alakhdar","year":"2024","journal-title":"J. Chem. Inf. Model."},{"key":"2026091409074041700_cit34","doi-asserted-by":"publisher","first-page":"5996","DOI":"10.1021\/acs.jcim.3c00667","article-title":"SILVR: Guided\nDiffusion for Molecule Generation","volume":"63","author":"Runcie","year":"2023","journal-title":"J. Chem.\nInf. Model."},{"key":"2026091409074041700_cit35","article-title":"LDMol: A text-to-molecule diffusion\nmodel with structurally informative latent space","author":"Chang","year":"2025","journal-title":"arXiv"},{"key":"2026091409074041700_cit36","doi-asserted-by":"publisher","first-page":"51","DOI":"10.1186\/s13321-020-00456-1","article-title":"An open\nsource chemical structure curation pipeline using RDKit","volume":"12","author":"Bento","year":"2020","journal-title":"J. Cheminform."},{"key":"2026091409074041700_cit37","article-title":"BERT:\nPre-training of Deep Bidirectional\nTransformers for Language Understanding","author":"Devlin","year":"2019","journal-title":"arXiv"},{"key":"2026091409074041700_cit38","article-title":"Uni-Mol: A Universal\n3D Molecular Representation Learning Framework","author":"Zhou","year":"2023","journal-title":"ChemRxiv"},{"key":"2026091409074041700_cit39","article-title":"ChemBERTa: Large-Scale\nSelf-Supervised Pretraining for Molecular Property Prediction","author":"Chithrananda","year":"2020","journal-title":"arXiv"},{"key":"2026091409074041700_cit40","doi-asserted-by":"publisher","first-page":"5","DOI":"10.1023\/A:1010933404324","article-title":"Random Forests","volume":"45","author":"Breiman","year":"2001","journal-title":"Mach. Learn,"},{"key":"2026091409074041700_cit41","article-title":"XGBoost: A Scalable\nTree Boosting System","author":"Chen","year":"2016","journal-title":"arXiv"},{"key":"2026091409074041700_cit42","article-title":"CatBoost: Unbiased\nBoosting with Categorical Features","author":"Prokhorenkova","year":"2018","journal-title":"arXiv"},{"key":"2026091409074041700_cit43","doi-asserted-by":"publisher","first-page":"1256","DOI":"10.1038\/s42256-022-00580-7","article-title":"Large-scale\nchemical language representations capture molecular structure\nand properties","volume":"4","author":"Ross","year":"2022","journal-title":"Nat. Mach. Intell."},{"key":"2026091409074041700_cit44","doi-asserted-by":"publisher","first-page":"662","DOI":"10.1039\/D5DD00348B","article-title":"ChemBERTa-3: an open\nsource training framework for chemical foundation\nmodels","volume":"5","author":"Singh","year":"2026","journal-title":"Digital Discovery"}],"container-title":["Journal of Chemical Information\nand Modeling"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/pubs.acs.org\/doi\/pdf\/10.1021\/acs.jcim.6c01547","content-type":"application\/pdf","content-version":"vor","intended-application":"unspecified"},{"URL":"https:\/\/pubs.acs.org\/jcisd8\/article-pdf\/66\/17\/10737\/68100579\/acs.jcim.6c01547.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/pubs.acs.org\/jcisd8\/article-pdf\/66\/17\/10737\/68100579\/acs.jcim.6c01547.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,14]],"date-time":"2026-09-14T13:14:43Z","timestamp":1789391683000},"score":1,"resource":{"primary":{"URL":"https:\/\/pubs.acs.org\/jcisd8\/article\/66\/17\/10737\/5395682\/PKP-Diffmol-A-Physicochemical-Knowledge-Prompt"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,1]]},"references-count":44,"journal-issue":{"issue":"17","published-online":{"date-parts":[[2026,9,1]]},"published-print":{"date-parts":[[2026,9,14]]}},"URL":"https:\/\/doi.org\/10.1021\/acs.jcim.6c01547","relation":{},"ISSN":["1549-9596","1549-960X"],"issn-type":[{"value":"1549-9596","type":"print"},{"value":"1549-960X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,1]]}}}