Abstract
Human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) have revolutionized cardiovascular research. Abnormalities in Ca2+ transients have been evident in many cardiac disease models. We have shown earlier that, by exploiting computational machine learning methods, normal Ca2+ transients corresponding to healthy CMs can be distinguished from diseased CMs with abnormal transients. Here our aim was to study whether it is possible to separate different genetic cardiac diseases (CPVT, LQT, HCM) on the basis of Ca2+ transients using machine learning methods. Classification accuracies of up to 87% were obtained for these three diseases, indicating that Ca2+ transients are disease-specific. By including healthy controls in the classifications, the best classification accuracy obtained was still high: approximately 79%. In conclusion, we demonstrate as the proof of principle that the computational machine learning methodology appears to be a powerful means to accurately categorize iPSC-CMs and could provide effective methods for diagnostic purposes in the future.
Highlights
Induced pluripotent stem cell-derived[1] cardiomyocytes have enabled the study of various genetic cardiac diseases such as catecholaminergic polymorphic ventricular tachycardia (CPVT)[2,3,4,5,6,7,8,9], long QT syndrome (LQT)[10,11,12,13] and hypertrophic cardiomyopathy (HCM)[14,15,16], and all of these have revealed substantial abnormalities and diversity in Ca2+ cycling properties when compared with healthy controls
Studied cell lines included six CPVT lines generated from CPVT patients carrying cardiac ryanodine receptor (RyR2) mutations: four HCM cell lines generated from HCM patients carrying either α-tropomyosin (TPM1) or myosin-binding protein C (MYBPC3) mutations, two LQT type 1 cell lines generated from patients carrying potassium voltage-gated channel subfamily Q member 1 (KCNQ1) mutations, and one cell line generated from a healthy control individual
Some computational machine learning methods, random forests and the least square support vector machine with an radial basis function kernel (RBF) kernel, including the computation of Ca2+ transient peak variable values, were shown to be a powerful tool to accurately separate the Ca2+ transient signals of the three diseases – including LQT1, HCM, and CPVT – from each other and from control WT iPSC-CMs with high classification accuracies (79–88%). This strongly indicates the possibility of discriminating between genetic cardiac diseases using Ca2+ transient profiles recorded from iPSC-CMs with signal analysis and machine learning classification methods
Summary
Induced pluripotent stem cell-derived[1] cardiomyocytes (iPSC-CMs) have enabled the study of various genetic cardiac diseases such as catecholaminergic polymorphic ventricular tachycardia (CPVT)[2,3,4,5,6,7,8,9], long QT syndrome (LQT)[10,11,12,13] and hypertrophic cardiomyopathy (HCM)[14,15,16], and all of these have revealed substantial abnormalities and diversity in Ca2+ cycling properties when compared with healthy controls. We compared visually normal and abnormal Ca2+ transient signals and peak variables from three genetic cardiac diseases, including CPVT, an exercise-induced malignant arrhythmogenic disorder[4,9]; LQT type 1, an electric disorder of the heart that predisposes patients to arrhythmias and sudden cardiac death[13]; and HCM, a disorder that affects the structure of heart muscle tissue with increased risk of arrhythmias and progressive heart failure[16] This comparison revealed that these diseases can be distinguished from each other based on our previously reported peak variable analysis[17] computed from these Ca2+ signals. The classes were compared and classified in two main ways: first, normal signals of the controls compared separately to either normal or abnormal Ca2+ transient signals of the three diseases, and second, signals of the combined normal and abnormal signals of the controls compared to combined normal and abnormal signals of each of the three diseases, i.e. four classes in total
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have