CStreet

CStreet infers cell state trajectories from time-series single-cell RNA sequencing (scRNA-seq) data to reveal temporal dynamics and probable transitions between discrete cell states (e.g., clusters or types).

Key Features:

• Cell State Trajectory Topology Inference: Infers the topology of cell state trajectories, capturing complex biological pathways with multiple branches and starting points rather than focusing on individual-cell trajectories.
• Time-Series Integration: Constructs k-nearest neighbor connections between cells within each time point and across adjacent time points using time-series information.
• Probability Estimation: Estimates connection probabilities among cell states to provide a probabilistic interpretation of potential transitions.
• Visualization: Represents inferred trajectories using a force-directed graph that can depict multiple starting points and paths.
• Performance Validation: Compared with six commonly used cell state trajectory reconstruction methods on simulated and real datasets, demonstrating high accuracy and robustness.

Scientific Applications:

• Developmental Biology: Maps temporal progression and branching of cell states during development from time-series scRNA-seq data.
• Disease Progression: Identifies temporal changes and probable transitions in cell states associated with disease progression.
• Regenerative Medicine: Characterizes trajectories and branching relevant to regeneration and cell-state reprogramming.

Methodology:

Constructs k-nearest neighbor connections using time-series data within and across adjacent time points, estimates probabilities of these connections to infer topology of cell state trajectories, and visualizes the inferred trajectories with a force-directed graph.