Machine Learning Identifies Potential Dual-Target Compounds for Atopic Dermatitis
Machine learning–based modeling may accelerate the discovery of novel therapies for atopic dermatitis (AD) by identifying compounds that simultaneously activate key anti-inflammatory pathways, according to a recent computational and experimental study. The findings highlight a potential strategy targeting both the aryl hydrocarbon receptor (AHR) and nuclear factor erythroid 2-related factor 2 (NRF2) pathways.
AHR and NRF2 signaling are increasingly recognized as important regulators of skin barrier function and oxidative stress, making dual activation an attractive therapeutic approach. However, few known compounds target both pathways, and traditional screening methods are resource intensive. To address this, investigators developed machine learning models using molecular descriptors and structural fingerprints to predict dual agonists.
The models demonstrated strong predictive performance, with all achieving area under the curve (AUC) values above 0.86. The authors reported that the optimal AHR model achieved “an accuracy of 0.811 and an AUC of 0.878,” while the NRF2 model reached “an accuracy of 0.839 and an AUC of 0.907.” These results suggest reliable classification of candidate compounds.
Using these models, the study identified structural features associated with dual activity, including moderate hydrophobicity, limited alkyl chains, and highly conjugated molecular structures. Screening of natural compound libraries yielded more than 1000 potential dual agonists, with several chemical classes, such as flavones and furocoumarins, showing promise.
Experimental validation in keratinocyte (HaCaT) cells confirmed that selected compounds, including indirubin and imperatorin, functioned as AHR/NRF2 dual agonists. The authors noted that this approach “provides a robust predictive tool” and “may support the discovery of anti-AD agents.”
Reference
Zhen Y, Li Q, Hu X, et al. AHR/NRF2 Dual agonist prediction and natural compound screening based on machine learning: a new strategy for the treatment of atopic dermatitis. Int J Mol Sci. 2026;27(8):3530. doi:10.3390/ijms27083530
