Date: August 3, 2025

Publisher: Jinan University Integrated Media Center

Editor: Chang Kaili

Atrial fibrillation, the most common cardiac arrhythmia worldwide, affects approximately 1% to 2% of the global population. Female patients face not only higher risks of complications but also generally poorer treatment outcomes. Recently, a research team led by Associate Professor Bai Jieyun and Professor Zhang Xiaoshen from Jinan University, in collaboration with international institutions including the University of Auckland and Ghent University, published a groundbreaking study titled Digital twin for sex-specific identification of class III antiarrhythmic drugs based on in vitro measurements, computer models, and machine learning tools in the prestigious computational biology journal PLOS Computational Biology. This research innovatively integrates digital twin technology with machine learning algorithms, achieving for the first time sex-specific accurate identification of Class III antiarrhythmic drugs and paving a new path for personalized treatment of atrial fibrillation.

The research team established a digital twin model library of cardiomyocytes comprising 5,663 male and 6,184 female virtual prototypes. After calibration with in vitro experimental data, the team systematically simulated the effects of 12 clinically common antiarrhythmic drugs—including 6 Class III and 6 non-Class III agents—on action potentials and calcium transients in cardiomyocytes. By extracting 14 key biomarkers such as resting membrane potential, action potential duration, and calcium transient amplitude, the researchers utilized machine learning algorithms including Support Vector Machine (SVM) to develop a sex-specific classifier, successfully achieving precise identification of Class III antiarrhythmic drugs.

Key findings of the study include:

The sex-specific model significantly improved predictive accuracy in drug identification, with the SVM classifier achieving over 89% accuracy and an F1-score exceeding 87%, representing an approximately 7% improvement compared to non-sex-specific models;

Changes in resting membrane potential (ΔRMP), action potential amplitude (ΔAPA), action potential duration (APD), and the newly proposed action potential area variation (ΔAREA) were identified as core biomarkers for distinguishing sex-specific drug responses;

Lower levels of IK1, INa, and Ito ion currents in female cardiomyocytes may represent an important mechanism underlying observed sex differences in drug response.

The study confirms that incorporating sex-specific factors significantly enhances the accuracy of antiarrhythmic drug classification, providing a quantitative tool for clinical precision medicine. For instance, this model enables more accurate prediction of female patients' responses to medications such as amiodarone and sotalol, thereby reducing the risk of life-threatening arrhythmias like torsades de pointes. Furthermore, the established digital twin-machine learning framework can simulate drug effects on cardiomyocytes of different sexes, enabling evaluation of sex-specific efficacy of candidate drugs without relying extensively on clinical trials, substantially accelerating drug screening processes and reducing R&D costs.

The corresponding authors of this paper are Associate Professor Bai Jieyun and Director Lu Hua from Jinan University. The research received support from multiple funding sources, including the National Natural Science Foundation of China, the Natural Science Foundation of Guangdong Province, and the National Foreign Experts Program.

Paper link: https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1013154