CardioShift: Resolving cell-to-cell functional heterogeneity in hiPSC-derived cardiomyocytes from high-throughput single-cell imaging

High-content screening of human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) is commonly employed for disease modeling. However, standard pipelines utilize whole-well averages, which fail to capture cell-to-cell heterogeneity. Transitioning from well-averages to single-cell analysis reveals these hidden variations, but may inadvertently introduce pseudoreplication bias if cells are treated as independent data points.

The researchers have developed CardioShift, an open-source bioimage informatics framework, to map single-cell distributions while properly accounting for the nested variance structure. By applying this framework to 32,818 Ca2+ transients and 10,719 contractility-lipid profiles, the team discovered that cell-to-cell variability is the primary source of variance for most functional parameters. This finding demonstrates how traditional well averages can obscure condition-dependent phenotypes.

Moreover, unsupervised clustering using CardioShift identified condition-dependent shifts in functional subpopulations. One notable discovery was a distinct cellular state that accumulates lipids but maintains baseline contraction kinetics. This specific cellular state was not detected using either well- or single-cell averaging approaches.

In summary, CardioShift offers a robust analytical approach to uncover hidden single-cell distributions, thereby enhancing our understanding of cellular heterogeneity in hiPSC-CM models.