Development and application of an in vitro system for Cardiac Disease Modeling and High-Throughput Screening

This thesis aims to develop a workflow for both generating hiPSC-CMs and analyzing cardiac-specific readouts in a medium- to high-throughput manner. Two methods were developed: an optical flow-based quantification tool designed to extract detailed contraction data (such as frequency and absolute pressure changes) in a mediumthroughput setting, and a machine learning algorithm capable of distinguishing between normal and aberrant contraction patterns while providing interpretability of the decision making process. These methods enable automated bulk analysis of cardiomyocyte recordings, offering a platform to study contractile function in response to exogenous factors and ECM composition changes.