Multimodal graph-based deep learning for predicting heat of combustion as a key hazard indicator under GHS/CLP standards
- 1Worker’s Care & Digital Health Lab, Department of Future Technology, Korea University of Technology and Education
- 2Department of Digital Anti-aging Healthcare (BK21), Inje University
Abstract
We report a multimodal graph-based deep learning framework that integrates molecular graph topology with physicochemical descriptors to predict the heat of combustion (HoC) directly from chemical structure. The model employs a neural network convolution (NNConv) to capture bonding and topological information, while molecular descriptors provide complementary global features. Trained on 4,516 experimental data points, the model achieved a mean absolute error (MAE) of 351.4 kJ/mol (1.89 kJ/g), corresponding to approximately 6.35% of the dataset mean (5529.65 kJ/mol), and a coefficient of determination (R²) of 0.942 on the test set. When benchmarked against linear regression, random forest, and graph-only GNN models, the corresponding MAE and R² values were 376.0 and 0.925, 381.7 and 0.914, and 1608.8 and 0.484, respectively, highlighting the benefit of integrating descriptors with graph-based learning using NNConv. The predicted HoC values enable reliable classification with respect to the 20 kJ/g GHS/CLP decision threshold, except for compounds within an ambiguity band of ± 1.89 kJ/g. Residual analysis showed low bias and an approximately normal error distribution. These results demonstrate that multimodal graph-based learning can provide decision-relevant HoC predictions, supporting chemical hazard screening while reducing reliance on experimental measurements.
Related
- Multi-Scale MRI, Radiology Reports, and Blood Biomarker–Guided Multimodal Deep Learning for Predicting Postoperative Recurrence in Cervical Cancer: A Multicenter Study
- Systematic Modality Ablation of Multimodal Machine Learning for Predicting 24-Month Progression from Mild Cognitive Impairment to Alzheimer's Disease
- A multimodal machine learning model for predicting reproductive outcomes in the rare congenital hypogonadotropic hypogonadism after treatment.
- A novel multimodal model integrating CT-based deep learning and clinical information for predicting 1-year OS of first-line immunotherapy in ES-SCLC patients
- A multimodal deep learning radiomics nomogram integrating clinicopathological features for predicting pathological complete response after neoadjuvant therapy in triple-negative breast cancer: a multicenter retrospective study
BibTeX
@article{nnyigide2026multimodal,
title = {{Multimodal graph-based deep learning for predicting heat of combustion as a key hazard indicator under GHS/CLP standards}},
author = {Osita Sunday Nnyigide and Haewon Byeon and Uchenna Esther Okpete},
journal = {Scientific Reports},
year = {2026},
month = jul,
doi = {10.1038/s41598-026-60540-8},
url = {https://doi.org/10.1038/s41598-026-60540-8},
note = {CC BY},
}
RIS
TY - JOUR
TI - Multimodal graph-based deep learning for predicting heat of combustion as a key hazard indicator under GHS/CLP standards
AU - Osita Sunday Nnyigide
AU - Haewon Byeon
AU - Uchenna Esther Okpete
PY - 2026
DA - 2026/07/01
JO - Scientific Reports
DO - 10.1038/s41598-026-60540-8
UR - https://doi.org/10.1038/s41598-026-60540-8
AB - We report a multimodal graph-based deep learning framework that integrates molecular graph topology with physicochemical descriptors to predict the heat of combustion (HoC) directly from chemical structure. The model employs a neural network convolution (NNConv) to capture bonding and topological information, while molecular descriptors provide complementary global features. Trained on 4,516 experimental data points, the model achieved a mean absolute error (MAE) of 351.4 kJ/mol (1.89 kJ/g), corresponding to approximately 6.35% of the dataset mean (5529.65 kJ/mol), and a coefficient of determination (R²) of 0.942 on the test set. When benchmarked against linear regression, random forest, and graph-only GNN models, the corresponding MAE and R² values were 376.0 and 0.925, 381.7 and 0.914, and 1608.8 and 0.484, respectively, highlighting the benefit of integrating descriptors with graph-based learning using NNConv. The predicted HoC values enable reliable classification with respect to the 20 kJ/g GHS/CLP decision threshold, except for compounds within an ambiguity band of ± 1.89 kJ/g. Residual analysis showed low bias and an approximately normal error distribution. These results demonstrate that multimodal graph-based learning can provide decision-relevant HoC predictions, supporting chemical hazard screening while reducing reliance on experimental measurements.
SN - 2045-2322
ER -