Transition state dynamics during a stochastic fate choice

A stochastic differentiation programme shows stepwise separation of cell fate, with initiation of cell-type specific gene expression coupled to global transcriptome changes shared by all cells. Overall design: Single cell transcriptomes of disaggregated mound cells were determined using the Fluidigm C1 system. An average of around 3 million reads were generated from each single cell library. We excluded seven cells (marked in Supplementary file Dd_Mound_RawCounts.csv) based on the following criteria: total number of reads was lower than 5 x 10^5, percentage of mitochondrial reads was higher than 15%, or the data was missing a contiguous part of the transcriptome. We obtained a total of 116 cells, from 3 experimental replicates.