sctoolsrna 0.1.0

A workbench of general-purpose single-cell RNA-seq analysis tools from the Wagner Lab (UCSF)

scToolsRNA

A workbench of general-purpose single-cell RNA-seq analysis tools from the Wagner Lab (UCSF).

scToolsRNA collects utilities that have accumulated across lab projects into a single pip-installable package built on top of scanpy and AnnData. It covers preprocessing and QC, feature selection and dimensionality reduction, differential expression, kNN label transfer, trajectory analysis, network export, plotting, and I/O.

It is a companion to the lab's dataset-specific zmap-tools package: where zmap-tools is specialized for the Zebrafish Multi-Atlas Project, scToolsRNA holds the organism-agnostic building blocks meant for everyday use.

Installation

pip install sctoolsrna

or, from a checkout of this repository:

pip install .

All runtime dependencies (scanpy, anndata, scikit-learn, faiss-cpu, pydeseq2, scrublet, umap-learn, harmonypy, plotly, igraph, leidenalg, …) are installed automatically. Individual modules import heavier or optional dependencies lazily (inside the functions that use them), so import scToolsRNA stays fast and does not fail if a single optional system library is missing.

Usage

The distribution installs as sctoolsrna, but the import name is scToolsRNA:

import scanpy as sc
import scToolsRNA as sct

adata = sc.read_h5ad("my_data.h5ad")

# Feature selection + significant-PC estimation
sct.get_variable_genes(adata, top_n_genes=3000)
sct.get_sig_pcs(adata)

# kNN label transfer from a labeled reference
sct.transfer_labels_knn(
    adata_query,
    adata_ref,
    ref_label_col="cell_type",
    ref_basis="X_pca_harmony",
    query_basis="X_pca_harmony",
)

Every public function is also re-exported at the top level, so from scToolsRNA import get_variable_genes continues to work for existing code.

Modules

Module	Contents
preprocess	Barcode/mito/ribo/doublet filtering and sampling QC
dimensionality	Variable-gene (V-score) selection, covarying genes, significant-PC estimation
workflows	Convenience pipelines (raw → normalized → UMAP/Leiden)
diffexp	Pseudobulk pyDESeq2 contrasts, DEG tables, volcano/clustermap plots
classification	Train/predict per-cell classifiers (sklearn)
knn	Portable FAISS/scikit-learn k-nearest-neighbor search
labeltransfer	kNN label and continuous-value transfer between datasets
trajectories	Diffusion-pseudotime dynamic-gene detection and plotting
stitch	STITCH temporal graph construction and diagnostics
network	GraphML / Pajek export for Gephi
plotting	3D embeddings, UMAP animations, axis helpers
colormaps	Custom matplotlib colormaps
utils	Label smoothing, confusion matrices, stacked barplots
readwrite	STARsolo / alevin / alevin-fry / inDrops loaders, cell/gene metadata
sparse	Sparse-matrix helpers

License

See LICENSE.