Abstract: Characterization of induced pluripotent stem cells (iPSC) and quantification of early differentiation events pose significant challenges to reliably determine differentiation potential. Here, we describe PluripotencyScreen – a combination of cell-type specific DNA methylation signatures to distinguish pluripotent, endoderm, mesoderm, neuronal and non-neuronal ectoderm cells. Furthermore, these signatures can be used for deconvolution to estimate the composition of cell types. PluripotencyScreen was established with the computational framework CimpleG to derive epigenetic signatures, consisting of a few CG dinucleotides per cell-type, enabling targeted analysis with digital PCR or pyrosequencing. We employed six defined differentiation conditions and used publicly available trilineage differentiation kits to generate genome wide DNA methylation profiles of defined cell states. Based on these datasets, we determined a pluripotency score to monitor reprogramming and to characterize pluripotent state of iPSCs, and lineage-specific scores to monitor either directed differentiation of iPSC or undirected multilineage differentiation in embryoid bodies. These scores have been validated through pyrosequencing and digital PCR for iPSC lines of various laboratories. The signatures have also revealed differentiation impairments, e.g. after genetic knockout of PRDM8 and YAP1. We demonstrate optimization and validation steps of our methodology to develop this approach in an easily applicable tool for the scientific community. Taken together, the cell-type specific DNA methylation changes underlying PluripotencyScreen facilitate reliable and reproducible quality control of iPSC and their derivates.
