Synthetic Applicability Domain (SynAD): Navigating Chemical Space for Reliable AI‐Driven Reaction Prediction
Angewandte Chemie International Edition · 23.02.2026 · 65 (9)
DOI: 10.1002/anie.202523874 ↗Kurz gefasst
Fachartikel in der Zeitschrift Angewandte Chemie International Edition, online erschienen am 22.01.2026. Verfasst von Zhenzhi Tan und 3 weiteren Autor:innen. Beteiligte Einrichtung: Center of Basic Molecular Science, Department of Chemistry, Tsinghua University, Beijing, 100084, P.R. China. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 46 Literaturangaben. Gefördert von: Natural Science Foundation of China, National Key Research and Development Program of China, Haihe Laboratory of Sustainable Chemical Transformations.
Abstract (Originalsprache)
Organic synthetic chemistry has undergone a paradigm shift driven by breakthroughs in artificial intelligence (AI). Data‐driven methods help accelerate hypothesis evaluation and reduce experimental trial‐and‐error efforts. However, its practical utility is constrained by the out‐of‐distribution (OOD) issue, where predictions usually fail when extrapolating to unseen reactions with new catalysts, substrates, or conditions. Here, we introduce SynAD (synthetic applicability domain), a machine learning framework for assessing the predictive capability of AI models trained with existing data. SynAD combines descriptors with model‐adaptive distance metrics to automatically demarcate reliable and unreliable reactions. Validated on the Ullmann Ligand Dataset (ULD, >5000 reactions), SynAD a priori distinguishes predictable chemical space, resulting in a prediction accuracy of R 2 = 0.90 (at 12.3% coverage) from a baseline of R 2 = −0.21. This capacity to target reliable chemical space is consistently observed across 6 additional datasets. We also enable a SynAD score to quantify reaction class predictability, guiding experimental focus on OOD spaces. By defining model limits, SynAD provides a critical guardrail for chemists to trust AI, allocate resources strategically, and accelerate de novo discovery.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Center of Basic Molecular Science, Department of Chemistry, Tsinghua University, Beijing, 100084, P.R. China.
- Haihe Laboratory of Sustainable Chemical Transformations, Tianjin, 300192, P.R. China.
- Key Laboratory of Organosilicon Chemistry and Material Technology of Ministry of Education, Key Laboratory of Organosilicon Material Technology of Zhejiang Province, College of Material, Chemistry and Chemical Engineeri…
ORCID-Profile: Zhenzhi Tan, Qi Yang, Sanzhong Luo
Förderung
- Natural Science Foundation of China
- National Key Research and Development Program of China
- Haihe Laboratory of Sustainable Chemical Transformations
- Fundamental and Interdisciplinary Disciplines Breakthrough Plan of the Ministry of Education of China
- National Natural Science Foundation of China
Zitieren
APA
Tan, Z., Yang, Q., Zhang, L., & Luo, S. (2026). Synthetic Applicability Domain (SynAD): Navigating Chemical Space for Reliable AI‐Driven Reaction Prediction. Angewandte Chemie International Edition, 65(9). https://doi.org/10.1002/anie.202523874
Vancouver
Tan Z, Yang Q, Zhang L, Luo S. Synthetic Applicability Domain (SynAD): Navigating Chemical Space for Reliable AI‐Driven Reaction Prediction. Angewandte Chemie International Edition. 2026;65(9). doi:10.1002/anie.202523874
BibTeX
@article{Tan2026,
author = {Tan, Zhenzhi and Yang, Qi and Zhang, Long and Luo, Sanzhong},
title = {Synthetic Applicability Domain (SynAD): Navigating Chemical Space for Reliable AI‐Driven Reaction Prediction},
journal = {Angewandte Chemie International Edition},
year = {2026},
volume = {65},
number = {9},
doi = {10.1002/anie.202523874}
}
EUVIMED