v0.1.0 - Initial Release

sclsd v0.1.0 - Initial Release

Single-cell Latent State Dynamics (sclsd) - A deep learning framework for modeling single-cell differentiation trajectories using Waddington-like potential landscapes.

Installation

pip install sclsd

Quick Start

from sclsd import LSD, LSDConfig

# Initialize with your AnnData object
lsd = LSD(adata, LSDConfig(), device=device)
lsd.set_phylogeny(phylogeny, cluster_key="clusters")
lsd.set_prior_transition(prior_time_key="prior_time")
lsd.prepare_walks()
lsd.train(num_epochs=100)

# Get results
result = lsd.get_adata()

Features

• Learn Waddington-like potential landscapes from single-cell data
• Trajectory inference with neural ODEs
• Fate prediction via gradient flow dynamics
• Integration with scanpy ecosystem