EpiMogrify Models H3K4me3 Data to Identify Signaling Molecules that Improve Cell Fate Control and Maintenance

Abstract

The need to derive and culture diverse cell or tissue types in vitro has prompted investigations on how changes in culture conditions affect cell states. However, the identification of the optimal conditions (e.g., signaling molecules and growth factors) required to maintain cell types or convert between cell types remains a time-consuming task. Here, we developed EpiMogrify, an approach that leverages data from ∼100 human cell/tissue types available from ENCODE and Roadmap Epigenomics consortia to predict signaling molecules and factors that can either maintain cell identity or enhance directed differentiation (or cell conversion). EpiMogrify integrates protein-protein interaction network information with a model of the cell's epigenetic landscape based on H3K4me3 histone modifications. Using EpiMogrify-predicted factors for maintenance conditions, we were able to better potentiate the maintenance of astrocytes and cardiomyocytes in vitro. We report a significant increase in the efficiency of astrocyte and cardiomyocyte differentiation using EpiMogrify-predicted factors for conversion conditions.

Keywords: cell conversion; cell maintenance; computational method; data driven; predictions.