Genetic Algorithm For Fitting Cardiac Cell Biophysical Model Formulations.

Area of Science:

• Computational biology
• Cardiovascular research
• Biophysics

Background:

• Cardiac cell electrophysiology modeling requires fitting complex equations to experimental data.
• Traditional fitting methods like trial-and-error or gradient-based optimization struggle with non-convex objective functions common in cardiac models.
• Meta-heuristic methods offer robust solutions against local optima and saddle points.

Purpose of the Study:

• To develop and validate a genetic algorithm-based approach for fitting cardiac electrophysiology models.
• To accurately parameterize biophysical model formulations using experimental data from human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs).

Main Methods:

• Utilized a genetic algorithm for fitting adult cardiomyocyte biophysical model equations.
• Employed whole-cell patch clamp data for rapid delayed rectifier potassium current (I_Kr), transient outward potassium current (I_to), and hyperpolarization-activated current (I_f).
• Implemented a two-point crossover scheme with constrained parameter space for initial population and mutation.

Main Results:

• Achieved near-optimal fitting of ionic current equations to experimental data.
• Obtained high R² values: 0.9960±0.0007 for I_Kr, 0.9995±0.0002 for I_to, and 0.9974±0.0014 for I_f (n=5).
• Demonstrated the genetic algorithm's efficacy in parameterizing cardiac ionic currents.

Conclusions:

• The genetic algorithm provides a robust and accurate method for fitting cardiac electrophysiology models.
• This approach successfully models key ionic currents (I_Kr, I_to, I_f) in hiPSC-CMs.
• The findings support the use of meta-heuristic methods for complex biophysical model parameterization.