py-SCORPIUS

A pure-Python port of SCORPIUS (Cannoodt et al., bioRxiv 2016) for linear-trajectory inference in single-cell RNA-seq.

• AnnData-compatible (cells × genes)
• No rpy2 — pure NumPy / SciPy / scikit-learn
• Same function surface as the R workflow (reduce_dimensionality → infer_trajectory → extract_modules → gene_importances)
• Pseudotime Pearson = 0.989 vs R SCORPIUS on canonical fixture
• Low-dim space Procrustes = 0.999 vs R

Install

pip install pyscorpius
# Optional — for extract_modules:
pip install pyscorpius[modules]

Quick-start

import numpy as np
from pyscorpius import reduce_dimensionality, infer_trajectory

# expression: (n_cells × n_genes) — or pass an AnnData and use .X.toarray()
space = reduce_dimensionality(expression, dist="spearman", ndim=3)
traj  = infer_trajectory(space, k=4)
pseudotime = traj["time"]    # in [0, 1] per cell
curve      = traj["path"]    # smooth curve through `space`

Function map

Reproducing R results

# Run R reference under your R conda env
Rscript tests/r_reference_driver.R data/fixture_simdata.rds data/reference_output.json

# Run Python candidate
python tests/_run_candidate.py data/fixture_simdata.rds data/candidate_output.json

# Compare (see compare_R_vs_Python.ipynb)
pytest tests/test_exact_match.py -v

Achieved on the SCORPIUS-bundled simulated dataset (400 cells × 200 genes):

• Pseudotime Pearson: 0.989 (threshold 0.95)
• Low-dim space Procrustes: 0.999 (threshold 0.85)

Relationship to omicverse

Developed under the omicverse-rebuildr protocol. Reuses py-mclustR for the extract_modules Mclust step.

Citation

Cannoodt, R. et al. SCORPIUS improves trajectory inference and identifies novel modules in dendritic cell development. bioRxiv 079509 (2016).

License

MIT — matches upstream SCORPIUS.