The scRNA-seq raw data of U2OS-FUCCI cell cycle from the paper by Mahdessian et al, which are available at GEO with accession GSE146773. Then, we could follow the snakemake pipeline to perform scRNA-Seq data preparation and general analysis including filter, dimensionality reduction and clustering analysis. The processed data of U2OS-FUCCI cell cycle is in cell_cycle.h5ad.

The scEU-seq data of RPE1-FUCCI cell cycle can be extracted using dynamo’s CLI: dyn.sample_data.scEU_seq_rpe1() or from the paper by Battich et al, which are available at GEO with accession number GSE128365.

The scRNA-seq raw data of the human fibroblast cell cycle from the paper by Riba et al, which are available at GEO with accession GSE167609.

For the scRNA-seq for ~1k U2OS-FUCCI data, we can estimate the RNA velocity by scvelo or dynamo.

For the scEU-seq data for ~3k RPE1-FUCCI cells, dyn.sample_data.scEU_seq_rpe1() to acquire the processed data, which includes cell cycle clustering and RNA velocity.

For the scRNA-seq data for ~3k human fibroblasts, the annotated data velocity_anndata_human_fibroblast_DeepCycle_ISMARA.h5ad with the cell cycle phase and RNA velocity can be downloaded from Zenodo.

We can reconstruct the vector field based RNA velocity with dyn.vf.VectorField() by using dynamo. Then, we can calculate the divergence, curl, acceleration and curvature to perform the differential geometry analysis. We can also calculate the jacobian to perform genetic perturbation and inference gene regulatory interaction.

We can learn an analytical function of vector field from sparse single cell samples on the entire space robustly by vector_field_function. Then, we could simulate stochastic dynamics by solving the Langevin equation based analytical function and quantify the non-equilibrium landscape-flux of the cell cycle.

L. Zhu, J. Wang, Quantifying Landscape-Flux via Single-Cell Transcriptomics Uncovers the Underlying Mechanism of Cell Cycle. Adv. Sci. 2024, 2308879. https://doi.org/10.1002/advs.202308879