Fig. 1: Overview of the TIVelo pipeline.
From: TIVelo: RNA velocity estimation leveraging cluster-level trajectory inference
a Constructing PAGA cluster graph for the dataset. b Finding possible cluster by root/end score as origin node. c Selecting main path beginning from the origin node. d Assigning diffusion pseudotime for cells along the main path. e Extracting time series \(\tilde{u}\) and \(\tilde{s}\) of each gene according to pseudotime t. Here \(\tilde{u}\) and \(\tilde{s}\) are rescaled to sum to 1 for each gene. f Calculating orientation score for G genes within the dataset. Identifying expected direction of the main path and setting new origin node. g Assigning level and child node(s) for each node in the cluster graph. This facilitates directed trajectory inference (DTI) analysis. h Building directed nearest neighbor (dNN) for each cell. i. Inferring RNA velocity for each cell by dNN.
