Single-cell Type Order Parameters: cell fate coordinates for gene expression data¶

scTOP is a Python package for identifying cell types from single-cell RNA-sequencing data by projecting samples onto reference cell type bases. This method uses Hopfield-inspired order parameters to define the coordinates of cell fate space, where differentiation trajectories can be studied.

The paper explaining the theory and mathematical details of this method is scTOP: physics-inspired order parameters for cellular identification and visualization by Maria Yampolskaya, Michael J. Herriges, Laertis Ikonomou, Darrell Kotton, and Pankaj Mehta.

Contents:

Installation¶

Install via pip:

pip install sctop

Or from source:

git clone https://github.com/Emergent-Behaviors-in-Biology/scTOP.git
cd scTOP
pip install -e .

Quick Start¶

Basic Usage¶

Load a pre-made reference basis:

import sctop as top

# List available bases
print(top.list_available_bases())

# Load a basis
basis, metadata = top.load_basis(basis_key="MCKO legacy")

Process and score your data:

# Process raw count data
processed_sample = top.process(raw_counts)

# Score against basis
projections = top.score(basis, processed_sample)

Create a custom basis from your own data:

import anndata as ad

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

# Create basis with validation
results = top.create_basis(
    adata=adata,
    cell_type_column='cell_type',
    threshold=100,
    do_anova=False
)

# Access the basis and metrics
basis = results['basis']
print(f"Accuracy: {results['metrics']['accuracy']:.3f}")

Sources for Reference Databases¶

Dependencies¶

Core Dependencies¶

NumPy
Pandas
SciPy
AnnData
scikit-learn
tables (PyTables)

Visualization¶

Matplotlib
Seaborn

Performance¶

Numba (for fast rank computation)
tqdm (progress bars)

Optional¶

requests (for downloading remote bases)

Citation¶

If you use scTOP in your research, please cite:

[Citation information to be added]

License¶

scTOP is released under the MIT License.