Using A Genetic Algorithm To Parameterize A Mathematical Model Of A Porcine Ventricular Cardiomyocyte

Following myocardial infarction (MI), the heart undergoes heterogeneous remodeling, with specific changes in the region surrounding the infarct scar, the border zone (BZ), compared to the more remote myocardium. Post-MI cardiomyocytes’ electrophysiological behavior depends on whether cells are in the BZ or remote. In this study, we seek to understand these heterogeneous electrophysiological changes as well as observed cell-cell variability. We use a genetic algorithm to fit cell-specific mathematical models for individual ventricular cardiomyocytes (vCMs) from post-MI pigs. MI was induced by 120min balloon occlusion of the left anterior descending coronary artery. vCMs were isolated from the BZ and a remote region one month later. Membrane potential recorded via whole-cell patch clamp and calcium transients measured using epifluorescence microscopy were used to fit the mathematical models. In the absence of a published pig cardiomyocyte model, we used the Tomek et al. (2019) human vCM model, modified to remove currents known to be absent in pig vCMs. Populations of cell models were generated by varying maximal conductances across a wide range. These variants were evaluated for their ability to recapitulate the observed electrophysiological data. The strongest variants were propagated with minor mutations and the process repeated for approximately thirty generations. Once the full dataset has been processed, the solutions for each cell will be analyzed to determine which parameter changes are required to fit the human model to the pig experimental data. Comparing the solutions for remote and BZ cells provides a molecular-level view of the changes involved and may identify key currents and transporters. Examining the variation in solutions among regions and cells will help explain what drives the heterogeneity of electrophysiological behavior post-MI.