5.12. RNA velocity estimation

RNA velocity estimation was implemented using the scVelo python package (v0.2.4; https://github.com/theislab/scvelo) based on python version 3.9.7 (using conda environment, v4.10.3). We extracted spliced and unspliced reads using the standard velocyto pipeline (v0.17.17; https://github.com/velocyto‐team/velocyto.py), which exported a loompy file. The file was read into an AnnData object for downstream analysis using Scanpy (v1.8.1; https://github.com/theislab/scanpy). Here we referenced the recommended workflow of scvelo (http://velocyto.org). Our analysis procedures included four steps: (1) data pre‐processing, (2) implementing the dynamical model to determine the full transcriptional dynamics of splicing kinetics, (3) recovering the latent time based on transcriptional dynamics and (4) projecting the latent time onto the UMAP embedding that was exported from the Seurat analysis mentioned above.