Structure Prototype Learning for Federated Molecular Property Prediction under Scaffold Heterogeneity
Journal of Chemical Information and Modeling · 09.11.2026 · 66 (21) · 13870-13886
DOI: 10.1021/acs.jcim.6c02297 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Chemical Information and Modeling, online erschienen am 20.10.2026. Verfasst von Tingkai Sun und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science, University of Tsukuba, Ibaraki305-8577, Japan. Hauptthemen laut MEDLINE: Small Molecule Libraries. Die Arbeit nennt 34 Literaturangaben. Gefördert von: Japan Society for the Promotion of Science, Japan Science and Technology Agency.
Abstract (Originalsprache)
Federated molecular property prediction enables collaboration across private chemical libraries, but parameter averaging does not explicitly account for scaffold heterogeneity. FedSPL groups molecules using fixed descriptors and exchanges embedding prototypes alongside model parameters, enabling label-independent prototype matching while retaining supervised property prediction. Across five classification benchmarks and three heterogeneity levels, matched three-seed evaluation using masked, held-out server-model ROC-AUC showed higher mean final AUC than FedAvg in seven of 15 settings. Consistently positive differences on SIDER and mixed outcomes elsewhere indicate data set-dependent benefits, motivating validation-guided use of structural information in federated prototype learning.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Substanzen
Autor:innen und Institutionen
- Department of Computer Science, University of Tsukuba, Ibaraki305-8577, Japan.
ORCID-Profile: Tingkai Sun, Tetsuya Sakurai
Förderung
- Japan Society for the Promotion of Science
- Japan Science and Technology Agency
Zitieren
APA
Sun, T., Ye, X., & Sakurai, T. (2026). Structure Prototype Learning for Federated Molecular Property Prediction under Scaffold Heterogeneity. Journal of Chemical Information and Modeling, 66(21), 13870-13886. https://doi.org/10.1021/acs.jcim.6c02297
Vancouver
Sun T, Ye X, Sakurai T. Structure Prototype Learning for Federated Molecular Property Prediction under Scaffold Heterogeneity. Journal of Chemical Information and Modeling. 2026;66(21):13870-13886. doi:10.1021/acs.jcim.6c02297
BibTeX
@article{Sun2026,
author = {Sun, Tingkai and Ye, Xiucai and Sakurai, Tetsuya},
title = {Structure Prototype Learning for Federated Molecular Property Prediction under Scaffold Heterogeneity},
journal = {Journal of Chemical Information and Modeling},
year = {2026},
volume = {66},
number = {21},
pages = {13870--13886},
doi = {10.1021/acs.jcim.6c02297}
}
EUVIMED