Empirical modeling of 1JCH coupling constants using genetic programming
Planta Medica · 01.03.2026 · 92 (03) · 295-296
DOI: 10.1055/s-0045-1815122 ↗Kurz gefasst
Fachartikel in der Zeitschrift Planta Medica, erschienen am 01.03.2026. Verfasst von M Aliberti und 3 weiteren Autor:innen.
Abstract (Originalsprache)
One-bond heteronuclear coupling constants ( 1 J CH ) are valuable for the configurational and conformational analysis of organic molecules. [ 1 ] [ 2 ] [ 3 ] For instance, they are employed to differentiate α- and β-anomeric protons in carbohydrates and to distinguish equatorial from axial protons in cyclohexane rings via the Perlin effect. However, the prediction of these NMR parameters by Density Functional Theory (DFT), useful for comparing them with experimental values, is limited by the high computational cost. To offer a more accessible alternative, we extended a previously developed computer-aided strategy, originally applied to the prediction of one-bond carbon-carbon coupling constants ( 1 J CC ), to the empirical prediction of 1 J CH values. Using genetic programming ( via gplearn) and scikit-learn [ 4 ], we generated mathematical models capable of accurately predicting experimental scalar coupling constants with reasonable computational costs and times. In this case, a dataset of 148 1 J CH values from diverse organic molecules was used to train the models. To enhance predictive performance, we integrated both NMR and structural descriptors, e.g., chemical shifts and bond angles. This multi-parametric approach significantly improved accuracy, achieving an R 2 value of 0.96, while greatly reducing computational cost compared to DFT. Beyond performance, the method is fast and interpretable, making it suitable for integration into stereochemical elucidation workflows. Future efforts will focus on validating this model on independent test sets. Additionally, comparisons among the values predicted by our developed empirical strategy, those DFT-predicted and experimental ones, will assess accuracy and generalizability. This approach will also be tested on selected case studies to further verify its utility in practical applications. Funding This research was supported by PRIN 2022 “2022RCTY7M – BiTMaP – Coupling the power of bioinformatics and experimental tools for accelerating the production, structural elucidation, and target identification of marine natural products”, CUP D53D23010460006, missione 4 “Istruzione e ricerca” Componente 2 “Dalla ricerca all’impresa” – investimento 1.1 del Piano Nazionale di Ripresa e Resilienza (PNRR) finanziato dall’Unione Europea “Next Generation EU”. Publication History Article published online: 05 March 2026 © 2026. Thieme. All rights reserved. Georg Thieme Verlag KG Oswald-Hesse-Straße 50, 70469 Stuttgart, Germany
Zitieren
APA
Aliberti, M., Chini, M. G., Bifulco, G., & Lauro, G. (2026). Empirical modeling of 1JCH coupling constants using genetic programming. Planta Medica, 92(03), 295-296. https://doi.org/10.1055/s-0045-1815122
Vancouver
Aliberti M, Chini MG, Bifulco G, Lauro G. Empirical modeling of 1JCH coupling constants using genetic programming. Planta Medica. 2026;92(03):295-296. doi:10.1055/s-0045-1815122
BibTeX
@article{Aliberti2026,
author = {Aliberti, M and Chini, M G and Bifulco, G and Lauro, G},
title = {Empirical modeling of 1JCH coupling constants using genetic programming},
journal = {Planta Medica},
year = {2026},
volume = {92},
number = {03},
pages = {295--296},
doi = {10.1055/s-0045-1815122}
}
EUVIMED