EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Toward Imbalanced Molecular Property Regression: A Benchmark Study and Interval-Aware Mixture of Experts

Yan Sun, Yu Shi, Alana Deng, Zihao Jing, Carson Leung iD, Pingzhao Hu iD

Journal of Chemical Information and Modeling · 2026

DOI: 10.1021/acs.jcim.6c01847 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Chemical Information and Modeling, 2026. Verfasst von Yan Sun und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Biochemistry, Western University, 1400 Western Road, London, ONN6G 2 V4, Canada. Hauptthemen laut MEDLINE: Drug Discovery. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 26 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: CIHR, Natural Sciences and Engineering Research Council of Canada, Canada Foundation for Innovation.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Molecular property prediction is a key task in AI-driven drug discovery, yet the prevalence and impact of label imbalances in molecular property regression remain poorly understood. Through a systematic benchmark of widely used molecular property data sets, we show that target values are often highly imbalanced and that prediction errors are consistently concentrated in sparsely represented regions of the label space. We further evaluate representative imbalance-learning approaches developed for general regression tasks and find that their effectiveness on molecular data sets is limited. To address this challenge, we propose interval-aware mixture-of-experts (IA-MoE), a plug-in framework that partitions the continuous target space into intervals and promotes expert specialization across different regions of the label distribution. IA-MoE can be seamlessly integrated with diverse molecular encoders without modifying their underlying architectures. Experiments on five molecular property benchmarks and four graph neural network backbones demonstrate that IA-MoE consistently improves the predictive performance in underrepresented regions while maintaining or improving overall accuracy. Our findings establish label imbalance as an important challenge in molecular property regression and highlight interval-aware expert learning as an effective strategy for addressing this challenge.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Drug DiscoveryBenchmarkingRegression AnalysisGraph Neural NetworksPrediction Algorithms

Autor:innen und Institutionen

  • Department of Biochemistry, Western University, 1400 Western Road, London, ONN6G 2 V4, Canada.
  • Department of Computer Science, Western University, 1151 Richmond Street, London, ONN6A 3K7, Canada.
  • Biostatistics Division, Dalla Lana School of Public Health, University of Toronto, 155 College Street, Toronto, ONM5T 3M7, Canada.
  • Department of Statistics and Actuarial Science, Simon Fraser University, Burnaby, BCV5A 1S6, Canada.
  • Department of Computer Science, University of Manitoba, Winnipeg, MBR3T 2N2, Canada.

ORCID-Profile: Carson Leung, Pingzhao Hu

Förderung

  • CIHR
  • Natural Sciences and Engineering Research Council of Canada
  • Canada Foundation for Innovation
  • Canada Research Chairs
  • Canadian Institutes of Health Research

Zitieren

APA

Sun, Y., Shi, Y., Deng, A., Jing, Z., Leung, C., & Hu, P. (2026). Toward Imbalanced Molecular Property Regression: A Benchmark Study and Interval-Aware Mixture of Experts. Journal of Chemical Information and Modeling. https://doi.org/10.1021/acs.jcim.6c01847

Vancouver

Sun Y, Shi Y, Deng A, Jing Z, Leung C, Hu P. Toward Imbalanced Molecular Property Regression: A Benchmark Study and Interval-Aware Mixture of Experts. Journal of Chemical Information and Modeling. 2026. doi:10.1021/acs.jcim.6c01847

BibTeX
@article{Sun2026,
  author = {Sun, Yan and Shi, Yu and Deng, Alana and Jing, Zihao and Leung, Carson and Hu, Pingzhao},
  title = {Toward Imbalanced Molecular Property Regression: A Benchmark Study and Interval-Aware Mixture of Experts},
  journal = {Journal of Chemical Information and Modeling},
  year = {2026},
  doi = {10.1021/acs.jcim.6c01847}
}
Kostenlos starten →