Figure 2.

Detecting the signal of cell fate commitment

The SGE value was calculated from Equation (5) based on MEF-to-neuron data (A), NPC-to-neuron data (B), hESC-to-DEC data (C), MHC-to-HCC data (D), and mESC-to-MP data (E), respectively. For each dataset, the significant increase of the SGE value (as shown as the red curve) indicates the imminent cell fate transition. Based on the top 5% genes with the highest and lowest local SGE values at the identified tipping point, t-SNE is applied to clustering cells for MEF-to-neuron data (F), NPC-to-neuron data (G), hESC-to-DEC data (H), MHC-to-HCC data (D), and mESC-to-MP data (E), respectively. Nodes in different colors represent cells from different time points. MEF, mouse embryonic fibroblast; NPC, neural progenitor cell; hESC, human embryonic stem cell; DEC, definitive endoderm cell; MHC, mouse hepatoblast cell; HCC, hepatocyte and cholangiocyte cell; mESC, mouse embryonic stem cell; MP, mesoderm progenitor; t-SNE, t-distributed stochastic neighbor embedding.