Pseudotime analysis

Pseudotime trajectories allow to order single cells according to a biological process of interest and to identify key paths or branches corresponding to alternative cellular states. We analyzed single cell trajectories using the Monocle2 algorithm (monocle R package v.2.12.0)^24,68,69. Starting from high dimensional data, this method projects cells into a lower dimensional space by constructing a principal graph and ordering them according to a pseudotime trajectory. Similarly to Cacchiarelli et al.⁷⁰, a set of ordering genes was defined as the 500 most DEGs over time, identified using the differentialGeneTest R function. The DDRTree method was used to project the cells into a two-dimensional space based on the expression of the selected genes, and simultaneously learn a graph structure into this space⁶⁹. Pseudotime values were then estimated as the distance of each cell from the root of the graph, which is defined as the state with the highest number of cells sampled at day 0 of differentiation. Finally, the cell-wise scaled pseudotime was computed dividing the estimated values by the maximum across all cells (Supplementary Data ²).