veloviz

VeloViz uses each cell's current observed and predicted future transcriptional states inferred from RNA velocity analysis to build a nearest neighbor graph between cells in the population, with edges pruned based on a cosine correlation threshold and/or a distance threshold and the resulting graph visualized using a force-directed graph layout algorithm to ensure relationships between cell states are reflected in the 2D embedding for more reliable representation of underlying cellular trajectories.

Key Features:

• RNA velocity-informed graph construction: Uses each cell's current observed and predicted future transcriptional states inferred from RNA velocity analysis to build a nearest neighbor graph between cells in the population.
• Edge pruning: Prunes edges based on a cosine correlation threshold and/or a distance threshold.
• Force-directed layout visualization: Visualizes the resulting graph using a force-directed graph layout algorithm.
• Trajectory-preserving embedding: Helps ensure that relationships between cell states are reflected in the 2D embedding for more reliable representation of underlying cellular trajectories.

Scientific Applications:

• Single-cell trajectory visualization: Generates 2D embeddings that reliably represent underlying cellular trajectories from RNA velocity data.
• Cell state relationship mapping: Enables visualization of relationships between cell states in a population based on current and predicted future transcriptional states.

Methodology:

VeloViz builds a nearest neighbor graph between cells using current observed and predicted future transcriptional states from RNA velocity analysis, prunes edges based on cosine correlation threshold and/or distance threshold, and applies a force-directed graph layout algorithm for visualization.