Fig. 3
From: Pumping the brakes on RNA velocity by understanding and interpreting RNA velocity estimates
The challenges in visualizing RNA velocity in embeddings. a Example of a gene g in the simulated data. From left to right, we show the phase portrait, spliced counts over true latent time, unspliced counts over true latent time, Ms over true latent time, Mu over true latent time, estimated velocity over true latent time, and comparisons of estimated velocity and true velocity. b True RNA velocity vector field on PCA using the transition probability method. c As b, but the high-dimensional velocities are mapped by the projection method. d Comparison of the low-dimensional speed produced by two methods in b and c. e Comparison between the low-dimensional speed produced by projection and the high-dimensional speed. f Comparison between the low-dimensional speed produced by transition probability and the high-dimensional speed (abbreviations: trans.pro, transition probability; high.dim, high-dimensional)