Application of the VelocityKernel

Application of the VelocityKernel#

In this analysis, we infer cell-cell transition probabilities using the VelocityKernel and show how to combine two kernels.

The corresponding data can be generated through the preprocessing notebooks velocity/preprocessing.ipynb, or downloaded driectly from here and should be saved in data/spermatogenesis/processed/adata.h5ad, i.e.,

mkdir -p ../../data/spermatogenesis/processed/
wget https://figshare.com/ndownloader/files/53395040 -O ../../data/spermatogenesis/processed/adata.h5ad

Library imports#

import anndata as ad
import cellrank as cr

from crp import DATA_DIR, FIG_DIR

General settings#

cr.settings.verbosity = 4

Constants#

SAVE_FIGURES = False
if SAVE_FIGURES:
    (FIG_DIR / "velocitykernel").mkdir(parents=True, exist_ok=True)

FIGURE_FORMATE = "svg"

SAVE_RESULTS = True
if SAVE_RESULTS:
    (DATA_DIR / "spermatogenesis" / "results").mkdir(parents=True, exist_ok=True)

Data loading#

adata = ad.io.read_h5ad(DATA_DIR / "spermatogenesis" / "processed" / "adata.h5ad")
adata

Data preprocessing#

VelocityKernel#

Kernel#

vk = cr.kernels.VelocityKernel(adata)
vk.compute_transition_matrix()

ck = cr.kernels.ConnectivityKernel(adata).compute_transition_matrix()
combined_kernel = 0.8 * vk + 0.2 * ck
Contents