Quick Start

This guide gives you the essential steps to get started with CellDyc.

Import CellDyc

import celldyc as cdc

Load the Data

Built‑in Dataset:

adata = cdc.datasets.celegans()  # 3 sampling time points, 333 cells

Your Own Dataset:

import scanpy as sc
adata = sc.read_h5ad("my_data.h5ad")  # requires .obs["time_point"] with discrete time labels

Preprocess the Data

One-liner to filter, normalize, pick HVGs, run PCA, and build the neighbor graph:

adata = cdc.tl.preprocess(adata)

Estimation of Time Representation and Transcriptomic Velocity

cdc.tl.recover_dyc(adata, time_key="time_point")

Outputs

  • adata.obs["getime"] – Continuous time representation (0‑1)

  • adata.layers["velocity"] – Velocity matrix

Visualization

  • cdc.pl.plot_velocity_projection(adata)

  • cdc.pl.getime_violin(adata, key_cont="getime", key_cat="time_point")

  • ……