{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,10]],"date-time":"2026-10-10T16:37:56Z","timestamp":1791650276543,"version":"4.3.4"},"reference-count":43,"publisher":"American Chemical Society (ACS)","issue":"14","license":[{"start":{"date-parts":[[2026,7,14]],"date-time":"2026-07-14T00:00:00Z","timestamp":1783987200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,7,14]],"date-time":"2026-07-14T00:00:00Z","timestamp":1783987200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"},{"start":{"date-parts":[[2026,7,14]],"date-time":"2026-07-14T00:00:00Z","timestamp":1783987200000},"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":["82003847"],"award-info":[{"award-number":["82003847"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["82574489"],"award-info":[{"award-number":["82574489"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Major Scientific and Technological Innovation Platform Project of Hunan Province","award":["2024JC1003"],"award-info":[{"award-number":["2024JC1003"]}]},{"name":"Science and Technology Development Program of Shanghai University of Traditional Chinese Medicine","award":["24KFL004"],"award-info":[{"award-number":["24KFL004"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,7,27]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Accurate prediction of molecular properties is essential for accelerating drug discovery, yet current deep learning methods generally lack reliable uncertainty estimates, particularly at the atomic level. A limitation of current approaches is the reliance on global molecular uncertainty, which frequently masks localized ambiguities by averaging the uncertainty across the entire molecule. To address this, we introduce AUINet, an atomic uncertainty-aware iterative network that quantifies and refines uncertainty at the atomic scale. Built on a D-MPNN architecture, AUINet uses Monte Carlo dropout to estimate atom-level uncertainty and iteratively refines atomic features through uncertainty-guided updates. Comprehensive evaluations show that AUINet outperforms state-of-the-art models on molecular property benchmarks and protein\u2013protein interaction inhibitor tasks under low-data conditions, all without requiring extensive pretraining. Crucially, rejection sampling experiments reveal that atom uncertainty provides a more robust error signal than classic molecular uncertainty. More importantly, AUINet provides chemically interpretable insights by pinpointing specific functional groups and structural motifs that contribute most to prediction uncertainty, as validated in solubility prediction and activity-cliff analysis. Overall, AUINet\u2019s precise localization of atomic-level uncertainty establishes a new paradigm for trustworthy molecular property prediction.<\/jats:p>","DOI":"10.1021\/acs.jcim.6c00973","type":"journal-article","created":{"date-parts":[[2026,7,14]],"date-time":"2026-07-14T17:58:53Z","timestamp":1784051933000},"page":"8076-8086","source":"Crossref","is-referenced-by-count":2,"title":["Atomic Uncertainty Pinpoints Critical Failure Structures\nfor Trustworthy Molecular Property Prediction"],"prefix":"10.1021","volume":"66","author":[{"ORCID":"https:\/\/orcid.org\/0009-0004-9929-3444","authenticated-orcid":true,"given":"Jiayu","family":"Qian","sequence":"first","affiliation":[{"name":"Central South University , , ,","place":["Changsha, People\u2019s Republic of China, 410083"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-6119-3149","authenticated-orcid":true,"given":"Yukun","family":"Luo","sequence":"additional","affiliation":[{"name":"Shanghai University of Traditional Chinese Medicine , , ,","place":["Shanghai, People\u2019s Republic of China, 201203"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Qingping","family":"Zhou","sequence":"additional","affiliation":[{"name":"Central South University , , ,","place":["Changsha, People\u2019s Republic of China, 410083"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xiaoqing","family":"Guan","sequence":"additional","affiliation":[{"name":"Shanghai University of Traditional Chinese Medicine , , ,","place":["Shanghai, People\u2019s Republic of China, 201203"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"316","published-online":{"date-parts":[[2026,7,14]]},"reference":[{"key":"2026072709093482300_cit1","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":"2026072709093482300_cit2","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-024-55082-4","article-title":"Multi-channel learning\nfor integrating structural hierarchies into context-dependent molecular\nrepresentation","volume":"16","author":"Wan","year":"2025","journal-title":"Nat. Commun."},{"key":"2026072709093482300_cit3","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-025-58503-0","article-title":"Uncertainty quantification with graph\nneural networks for efficient\nmolecular design","volume":"16","author":"Chen","year":"2025","journal-title":"Nat. Commun."},{"key":"2026072709093482300_cit4","unstructured":"Rong, Y.; et al. In  Self-Supervised\nGraph Transformer on Large-Scale Molecular Data, NeurIPS\n2020, 2020."},{"key":"2026072709093482300_cit5","doi-asserted-by":"publisher","first-page":"127","DOI":"10.1038\/s42256-021-00438-4","article-title":"Geometry-enhanced molecular representation\nlearning for property prediction","volume":"4","author":"Fang","year":"2022","journal-title":"Nat. Mach.\nIntell."},{"key":"2026072709093482300_cit6","doi-asserted-by":"crossref","unstructured":"Zhou, G.; Gao, Z.; Ding, Q.; et al. In  UNI-MOL: A Universal 3D Molecular\nRepresentation Learning Framework, Eleventh International\nConference on Learning Representations, 2023.","DOI":"10.26434\/chemrxiv-2022-jjm0j-v4"},{"key":"2026072709093482300_cit7","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-024-45461-2","article-title":"DynamicBind: predicting\nligand-specific protein-ligand complex structure with a deep equivariant\ngenerative model","volume":"15","author":"Lu","year":"2024","journal-title":"Nat. Commun."},{"key":"2026072709093482300_cit8","doi-asserted-by":"publisher","first-page":"10819","DOI":"10.1021\/acs.jcim.5c00464","article-title":"Upgrading\nreliability in molecular property prediction by robust quantification\nof uncertainty from machine learning models","volume":"65","author":"Ko\u0308tter","year":"2025","journal-title":"J. Chem. Inf. Model."},{"key":"2026072709093482300_cit9","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/ac3eb3","article-title":"Calibrated\nuncertainty\nfor molecular property prediction using ensembles of message passing\nneural networks","volume":"3","author":"Busk","year":"2022","journal-title":"Mach. Learn. Sci. Technol."},{"key":"2026072709093482300_cit10","doi-asserted-by":"publisher","first-page":"7699","DOI":"10.1021\/acs.jcim.3c01241","article-title":"An Uncertainty-Guided Deep Learning Method\nFacilitates Rapid Screening of CYP3A4 Inhibitors","volume":"63","author":"Wang","year":"2023","journal-title":"J. Chem. Inf. Model."},{"key":"2026072709093482300_cit11","doi-asserted-by":"publisher","first-page":"8131","DOI":"10.1021\/acs.jcim.4c01358","article-title":"Uncertainty qualification for deep\nlearning-based elementary\nreaction property prediction","volume":"64","author":"Liu","year":"2024","journal-title":"J. Chem. Inf.\nModel."},{"key":"2026072709093482300_cit12","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-025-62235-6","article-title":"Evidential deep learning-based\ndrug-target interaction prediction","volume":"16","author":"Zhao","year":"2025","journal-title":"Nat. Commun."},{"key":"2026072709093482300_cit13","doi-asserted-by":"publisher","first-page":"243","DOI":"10.1016\/j.inffus.2021.05.008","article-title":"A review of uncertainty\nquantification in deep learning: Techniques, applications and challenges","volume":"76","author":"Abdar","year":"2021","journal-title":"Inf. Fusion"},{"key":"2026072709093482300_cit14","unstructured":"Lakshminarayanan, B.; Pritzel, A.; Blundell, C. In  Simple\nand Scalable Predictive Uncertainty Estimation using Deep Ensembles, NeurIPS 2017, 2017."},{"key":"2026072709093482300_cit15","unstructured":"Guo, C.; Pleiss, G.; Sun, Y.; Weinberger, K. Q.\n          On Calibration of Modern Neural\nNetworks, 2017. 10.48550\/arXiv.1706.04599."},{"key":"2026072709093482300_cit16","doi-asserted-by":"crossref","unstructured":"Mukhoti, J.; Kirsch, A.; Van Amersfoort, J.; Torr, P. H. S.; Gal, Y.In  Deep Deterministic Uncertainty:\nA New Simple Baseline, IEEE\/CVF Conference on Computer Vision\nand Pattern Recognition (CVPR), 2023.","DOI":"10.1109\/CVPR52729.2023.02336"},{"key":"2026072709093482300_cit17","unstructured":"Amini, A.; Schwarting, W.; Soleimany, A.In  Deep Evidential Regression, NeurIPS 2020, 2020."},{"key":"2026072709093482300_cit18","doi-asserted-by":"publisher","first-page":"1076","DOI":"10.1016\/j.ijrobp.2021.12.011","article-title":"Understanding the differences\nbetween bayesian and frequentist statistics","volume":"112","author":"Fornacon-Wood","year":"2022","journal-title":"Int. J. Radiat. Oncol. Biol. Phys."},{"key":"2026072709093482300_cit19","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-022-34025-x","article-title":"Uncertainty-informed\ndeep learning models enable high-confidence predictions for digital\nhistopathology","volume":"13","author":"Dolezal","year":"2022","journal-title":"Nat. Commun."},{"key":"2026072709093482300_cit20","unstructured":"Blundell, C.; Cornebise, J.; Kavukcuoglu, K.; Wierstra, D.\n          Weight Uncertainty\nin Neural Networks, 2015. 10.48550\/arXiv.1505.05424."},{"key":"2026072709093482300_cit21","unstructured":"Gal, Y.; Ghahramani, Z.\n          Dropout as a Bayesian\nApproximation: Representing Model Uncertainty in Deep Learning, 2016."},{"key":"2026072709093482300_cit22","unstructured":"Damianou, A. C.; Lawrence, N. D.\n          Deep Gaussian\nProcesses, 2013. 10.48550\/arXiv.1211.0358."},{"key":"2026072709093482300_cit23","unstructured":"Hu, W.; et al.\n          Strategies\nfor Pre-training Graph Neural Networks, 2020. 10.48550\/arXiv.1905.12265."},{"key":"2026072709093482300_cit24","doi-asserted-by":"publisher","first-page":"279","DOI":"10.1038\/s42256-022-00447-x","article-title":"Molecular Contrastive\nLearning of Representations via Graph Neural Networks","volume":"4","author":"Wang","year":"2022","journal-title":"Nat. Mach. Intell."},{"key":"2026072709093482300_cit25","unstructured":"Liu, S.; et al.\n          Pre-training\nMolecular Graph Representation with 3D Geometry, 2022. 10.48550\/arXiv.2110.07728."},{"key":"2026072709093482300_cit26","doi-asserted-by":"crossref","unstructured":"Hou, Z.; Rangwala, H.; Hou, Z.; et al.In  GraphMAE: Self-Supervised\nMasked Graph Autoencoders, 28th ACM SIGKDD Conference on\nKnowledge Discovery and Data Mining, 2022; pp 594\u2013604.","DOI":"10.1145\/3534678.3539321"},{"key":"2026072709093482300_cit27","doi-asserted-by":"publisher","first-page":"1004","DOI":"10.1038\/s42256-022-00557-6","article-title":"Accurate prediction\nof molecular properties and drug targets using a self-supervised image\nrepresentation learning framework","volume":"4","author":"Zeng","year":"2022","journal-title":"Nat. Mach.\nIntell."},{"key":"2026072709093482300_cit28","doi-asserted-by":"publisher","first-page":"1256","DOI":"10.1038\/s42256-022-00580-7","article-title":"Large-scale chemical\nlanguage representations capture molecular structure and properties","volume":"4","author":"Ross","year":"2022","journal-title":"Nat. Mach. Intell."},{"key":"2026072709093482300_cit29","doi-asserted-by":"crossref","unstructured":"Yang, J.; Zheng, K.; Long, S.; et al.\n          Mol-AE: Auto-Encoder\nBased Molecular Representation Learning With 3D Cloze Test Objective, 2024. 10.1101\/2024.04.13.589331.","DOI":"10.1101\/2024.04.13.589331"},{"key":"2026072709093482300_cit30","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-023-43214-1","article-title":"A knowledge-guided pre-training\nframework for improving molecular representation learning","volume":"14","author":"Li","year":"2023","journal-title":"Nat. Commun."},{"key":"2026072709093482300_cit31","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-025-59634-0","article-title":"A self-conformation-aware pre-training framework\nfor molecular property prediction with substructure interpretability","volume":"16","author":"Qiao","year":"2025","journal-title":"Nat. Commun."},{"key":"2026072709093482300_cit32","doi-asserted-by":"publisher","first-page":"435","DOI":"10.1146\/annurev-pharmtox-011613-140028","article-title":"Targeting\nProtein-Protein Interaction\nby Small Molecules","volume":"54","author":"Jin","year":"2014","journal-title":"Annu. Rev. Pharmacol. Toxicol."},{"key":"2026072709093482300_cit33","doi-asserted-by":"publisher","first-page":"5438","DOI":"10.1021\/acs.jcim.1c01135","article-title":"pdCSM-PPI:\nUsing Graph-Based Signatures\nto Identify Protein\u2013Protein Interaction Inhibitors","volume":"61","author":"Rodrigues","year":"2021","journal-title":"J. Chem. Inf. Model."},{"key":"2026072709093482300_cit34","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiomed.2024.108287","article-title":"PPII-AEAT:\nPrediction\nof protein-protein interaction inhibitors based on autoencoders with\nadversarial training","volume":"172","author":"Zhang","year":"2024","journal-title":"Comput. Biol. Med."},{"key":"2026072709093482300_cit35","doi-asserted-by":"publisher","first-page":"274","DOI":"10.1021\/acs.jcim.1c01163","article-title":"Benchmarking\nMolecular Feature Attribution\nMethods with Activity Cliffs","volume":"62","author":"Jime\u0301nez-Luna","year":"2022","journal-title":"J. Chem. Inf.\nModel."},{"key":"2026072709093482300_cit36","doi-asserted-by":"publisher","first-page":"5938","DOI":"10.1021\/acs.jcim.2c01073","article-title":"Exposing the Limitations\nof Molecular Machine Learning with Activity Cliffs","volume":"62","author":"Van Tilborg","year":"2022","journal-title":"J. Chem. Inf. Model."},{"key":"2026072709093482300_cit37","doi-asserted-by":"publisher","first-page":"47","DOI":"10.1186\/s13321-023-00708-w","article-title":"Exploring QSAR models\nfor activity-cliff prediction","volume":"15","author":"Dablander","year":"2023","journal-title":"J. Cheminf."},{"key":"2026072709093482300_cit38","doi-asserted-by":"crossref","unstructured":"Shen, W.; Cui, C.; Su, X.; et al.\n          Activity Cliff-Informed\nContrastive Learning for Molecular Property Prediction, 2024. 10.21203\/rs.3.rs-2988283\/v2.","DOI":"10.21203\/rs.3.rs-2988283\/v2"},{"key":"2026072709093482300_cit39","doi-asserted-by":"publisher","first-page":"3370","DOI":"10.1021\/acs.jcim.9b00237","article-title":"Analyzing\nLearned Molecular\nRepresentations for Property Prediction","volume":"59","author":"Yang","year":"2019","journal-title":"J.\nChem. Inf. Model."},{"key":"2026072709093482300_cit40","unstructured":"Herna\u0301ndez-Lobato, J. M.; Adams, R. P.\n          Probabilistic Backpropagation for Scalable Learning of Bayesian Neural\nNetworks, 2015. 10.48550\/arXiv.1502.05336."},{"key":"2026072709093482300_cit41","doi-asserted-by":"publisher","first-page":"9","DOI":"10.1021\/acs.jcim.3c01250","article-title":"Chemprop:\nA Machine\nLearning Package for Chemical Property Prediction","volume":"64","author":"Heid","year":"2024","journal-title":"J. Chem. Inf. Model."},{"key":"2026072709093482300_cit42","doi-asserted-by":"crossref","unstructured":"Seabold, S.; Perktold, J.\n          Statsmodels: Econometric\nand Statistical Modeling with Python.  SCIPY\n2010, 2010, pp 92\u201396. 10.25080\/Majora-92bf1922-011.","DOI":"10.25080\/Majora-92bf1922-011"},{"key":"2026072709093482300_cit43","unstructured":"Adamczyk, J.; et al.\n          rdkit\/rdkit:\n2024_03_5 (Q1 2024) Release, 2024. 10.5281\/zenodo.12782092."}],"container-title":["Journal of Chemical Information and Modeling"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/pubs.acs.org\/doi\/pdf\/10.1021\/acs.jcim.6c00973","content-type":"application\/pdf","content-version":"vor","intended-application":"unspecified"},{"URL":"https:\/\/pubs.acs.org\/jcisd8\/article-pdf\/66\/14\/8076\/65921809\/acs.jcim.6c00973.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/pubs.acs.org\/jcisd8\/article-pdf\/66\/14\/8076\/65921809\/acs.jcim.6c00973.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,27]],"date-time":"2026-07-27T13:31:48Z","timestamp":1785159108000},"score":1,"resource":{"primary":{"URL":"https:\/\/pubs.acs.org\/jcisd8\/article\/66\/14\/8076\/5208841\/Atomic-Uncertainty-Pinpoints-Critical-Failure"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,14]]},"references-count":43,"journal-issue":{"issue":"14","published-online":{"date-parts":[[2026,7,14]]},"published-print":{"date-parts":[[2026,7,27]]}},"URL":"https:\/\/doi.org\/10.1021\/acs.jcim.6c00973","relation":{},"ISSN":["1549-9596","1549-960X"],"issn-type":[{"value":"1549-9596","type":"print"},{"value":"1549-960X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,14]]}}}