Detection of genetic cardiac diseases by Ca2+ transient profiles using machine learning methods

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.

Separation of HCM and LQT Cardiac Diseases with Machine Learning of Ca2+ Transient Profiles

“…Patient-specific iPSC lines were established and characterized as described earlier. 9 Studied cell lines included four HCM cell lines generated from HCM patients carrying either TPM1 or myosin-binding protein C (MYBPC3) mutations, and two LQT1 cell lines generated from patients carrying KCNQ1 mutations, and four LQT2 cell lines generated from patients carrying mutation in HERG. IPSCs were differentiated into spontaneously beating CMs with END2-differentiation method 17 and dissociated into single cell level for Ca 2þ imaging studies, which was conducted in spontaneously beating Fura-2 AM (Invitrogen, Molecular Probes) or Fluo-4 AM (Life Technologies Ltd)-loaded CMs as described earlier.…”

“…If a signal included only normally developed peaks, it was determined to be normal. Earlier we categorized them visually to the type of normal and abnormal transient signals, but, as mentioned above, after having discovered that they can be classified together, 6,9 we classified them with both types together in the present research. transient measurements has increased in time with the updated measuring system, which explains its variation in the gained data.…”

“…The peak interval attribute characterizes regularity or irregularity of CM cycling. Later 9,16 five additional attributes were taken: the maximum and absolute minimum of the second derivative values from the right peak side (left side not used typically being short), surface area given by the peak curve and the line from the peak beginning to its end, duration from the peak beginning to the location of the first derivative maximum within the peak left side, and duration from the location of the first derivative absolute minimum within the peak right side to the peak maximum. As new in the current research we employed two following attributes.…”

“…[5][6][7] Functional studies of CM calcium cycling have revealed new insights into different cardiac dis-eases by showing substantial defects and abnormalities in the calcium cycling of iPSC-CMs, reflecting the cardiac phenotype observed in patients. 1,8 Calcium cycling has a central role in cardiac functionality by linking electrical activation and contraction, and therefore it is important to investigate it to improve the studies of disease pathology, prevention, and treatment but also, as we have shown earlier, 9 in disease diagnostics.…”

“…Furthermore, it was observed that three genetic cardiac diseases including jointly HCM and LQT1 and catecholaminergic polymorphic ventricular tachycardia can be separated from healthy controls (wild type, WT) by classifying the Ca 2þ transient signals of the diseased as one class and those of the controls as the other. 16 It was showed that those three diseases can be separated from each other and each of them from the CMs of the healthy controls, 9 where it was also discovered that it was not necessary to consider abnormally and normally beating CMs within each group as separate groups. They could be classified jointly, since both groups could be separated approximately equally well between three diseases and controls.…”