monocle3-python 1.4.26.post1

Python port of the R monocle3 package (tracks cole-trapnell-lab/monocle3 1.4.26 @ 4f4239a) — AnnData-based single-cell trajectory, clustering, and differential expression.

monocle3-python

Python port of the R monocle3 single-cell trajectory toolkit.

The port keeps monocle3's orchestration and R-side numerical behaviour while replacing the cell_data_set S4 container with anndata.AnnData and delegating non-monocle-specific algorithms to the scverse ecosystem (umap-learn, openTSNE, leidenalg, statsmodels, …).

Installation

pip install monocle3-python                # from PyPI

For local development:

git clone https://github.com/Bio-Babel/Monocle3-python.git
cd Monocle3-python
pip install -e ".[dev]"

Quickstart

import monocle3 as m3

adata = m3.load_packer_embryo()                  # AnnData (6188 × 20222)
m3.preprocess_cds(adata, num_dim=50)              # size-factor norm + truncated PCA
m3.align_cds(adata, alignment_group="batch")
m3.reduce_dimension(adata)                        # UMAP via umap-learn
m3.cluster_cells(adata)                           # Leiden
m3.learn_graph(adata)                             # SimplePPT principal graph
m3.order_cells(adata, root_pr_nodes=["Y_1"])      # pseudotime
m3.plot_cells(adata, color_cells_by="pseudotime")

Tutorials

Runnable notebooks that reproduce the R monocle3 vignettes live under tutorials/:

Notebook	Dataset
c_elegans_embryo_v2.ipynb	Packer et al. 2019 — C. elegans embryo trajectory
c_elegans_L2_v2.ipynb	Cao et al. 2017 — C. elegans L2 clustering + Garnett markers

License

Artistic-2.0, matching the upstream R monocle3.