Applicability-Domain-Aware Evaluation of Machine-Learning NMR Chemical-Shift Prediction in Small drug Molecules: GT-NMR for 13C and CASCADE for 1H
Abstract
A small external evaluation was performed to examine the practical accuracy and applicability domain of two pretrained machine-learning models for solution-state NMR chemical-shift prediction in five pharmaceutically relevant small molecules: amylmetacresol, benzocaine, dopamine, betazole, and betahistine.The pretrained GT-NMR graphtransformer model was used for atom-specific 13C chemical-shift prediction, while the CASCADE 3D graph neural-network workflow was used for 1H chemical-shift prediction.Predictions were compared with primary-literature NMR data where numerical assignments could be verified.Exact dataset overlap, nearest-training-set Morgan-fingerprint Tanimoto similarity, Murcko-scaffold coverage, and molecular complexity were evaluated as applicability-domain diagnostics.None of the five exact structures was detected in the checked GT-NMR or CASCADE proton train/validation/test partitions, although all five Murcko scaffolds were represented in the corresponding training sets.For GT-NMR, audited 13C molecule-level MAEs were 0.50 ppm for benzocaine, 2.59 ppm for
Related
- A Machine Learning Approach to logBB Prediction Combining Molecular Docking and Quantum-Chemical Descriptors
- MSIonization: A Machine Learning Tool for Ionization Mode Prediction of Small Molecules.
- Representing chemical and enzymatic reactions in fingerprint space for applicability filtering and classification
- Blood–Brain Barrier Permeability Prediction by Integrating Deep Learning Representations with Chemical Compound Knowledge
- Machine learning prediction of bismuth recovery from literature data: publication-grouped validation, anchored calibration and chemical interpretation
BibTeX
@misc{zadeh2026applicabilitydomainaware,
title = {{Applicability-Domain-Aware Evaluation of Machine-Learning NMR Chemical-Shift Prediction in Small drug Molecules: GT-NMR for 13C and CASCADE for 1H}},
author = {Diba Allameh Zadeh},
howpublished = {ChemRxiv},
year = {2026},
month = sep,
doi = {10.26434/chemrxiv.15008562/v1},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008562/v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Applicability-Domain-Aware Evaluation of Machine-Learning NMR Chemical-Shift Prediction in Small drug Molecules: GT-NMR for 13C and CASCADE for 1H
AU - Diba Allameh Zadeh
PY - 2026
DA - 2026/09/09
PB - ChemRxiv
DO - 10.26434/chemrxiv.15008562/v1
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008562/v1
AB - A small external evaluation was performed to examine the practical accuracy and applicability domain of two pretrained machine-learning models for solution-state NMR chemical-shift prediction in five pharmaceutically relevant small molecules: amylmetacresol, benzocaine, dopamine, betazole, and betahistine.The pretrained GT-NMR graphtransformer model was used for atom-specific 13C chemical-shift prediction, while the CASCADE 3D graph neural-network workflow was used for 1H chemical-shift prediction.Predictions were compared with primary-literature NMR data where numerical assignments could be verified.Exact dataset overlap, nearest-training-set Morgan-fingerprint Tanimoto similarity, Murcko-scaffold coverage, and molecular complexity were evaluated as applicability-domain diagnostics.None of the five exact structures was detected in the checked GT-NMR or CASCADE proton train/validation/test partitions, although all five Murcko scaffolds were represented in the corresponding training sets.For GT-NMR, audited 13C molecule-level MAEs were 0.50 ppm for benzocaine, 2.59 ppm for
ER -