The analysis is part of a project to evaluate the predictive power of electrophysiological models of cardiomyocytes. Cardiomyocyte models are commonly used to generate populations of models that hypothesize the electrical behavior of actual cardiac cells. I am interested in understanding if a single model represents one specific cardiac cell or if a subset of the population of models approximates a single cell.
FILES:
The file 'cluster_tools.py' is a module that contains clustering functions with default settings and plotting functions. It can be installed to whichever environment suits you best, but because it is quite generic I usually end up placing it in my 'site-packages' directory.
The model data sets used here represent a population of models that have been trained using an evolutionary algorithm (EA) to single-cell data set. The EA was independently run multiple times (at least 5x) and the 10% of models with the best EA scores were saved from each EA run. The resulting data set is pooled.