py-spacetrooper 0.1.0

Quality Control for image-based spatial transcriptomics data — Python port of SpaceTrooper

py-spacetrooper

Python port of SpaceTrooper — Quality Control for image-based spatial transcriptomics data.

Install

pip install py-spacetrooper

Quickstart

import anndata as ad
from spacetrooper import SpaceTrooper

# Load your spatial data
adata = ad.read_h5ad("your_data.h5ad")

# Run QC pipeline
st = SpaceTrooper(adata)
st.spatial_per_cell_qc()
st.compute_qc_score()

# Access results
print(st.adata.obs["QC_score"])  # QC scores (0-1)

Functional API

from spacetrooper import (
    spatial_per_cell_qc,
    compute_outliers_qc_score,
    check_outliers,
    compute_qc_score,
)

spatial_per_cell_qc(adata)
compute_outliers_qc_score(adata)
check_outliers(adata)
compute_qc_score(adata)

What's included

Python function	R equivalent	Description
spatial_per_cell_qc	spatialPerCellQC	Per-cell QC metrics
compute_spatial_outlier	computeSpatialOutlier	Outlier detection (medcouple/MAD)
compute_outliers_qc_score	computeOutliersQCScore	Outlier labels for QC formula
check_outliers	checkOutliers	Validate outlier counts
compute_qc_score	computeQCScore	Ridge logistic regression QC score
compute_threshold_flags	computeThresholdFlags	Fixed threshold flags
read_cosmx_spe	readCosmxSPE	Read CosMx data
read_merfish_spe	readMerfishSPE	Read MERFISH data
read_xenium_spe	readXeniumSPE	Read Xenium data
read_polygons	readPolygons	Read polygon boundaries
plot_metric_hist	plotMetricHist	Metric histogram
plot_centroids	plotCentroids	Spatial scatter plot

Parity with R

• Pearson r = 0.989 against R computeQCScore output
• Intermediate metrics match to machine precision (< 1e-13)
• 102× faster than R (0.068s vs 6.956s on 905-cell CosMx data)

Citation

@article{righelli2024spacetrooper,
  title={SpaceTrooper: Quality Control for image-based spatial transcriptomics},
  author={Righelli, Dario and others},
  year={2024}
}

License

MIT (matches upstream SpaceTrooper)