CellRank

CellRank computes probabilistic fate maps from single-cell RNA-seq data by integrating trajectory inference with RNA velocity to predict initial, intermediate, and terminal cell populations and their fate potentials across contexts including regeneration and disease.

Key Features:

• Integration of Trajectory Inference and RNA Velocity: Calculates RNA velocity from ratios of spliced to unspliced mRNA reads and integrates these directional signals with trajectory inference to predict future cellular states.
• Handling Stochasticity and Uncertainty: Models and incorporates uncertainty in RNA velocity vectors to account for stochastic and gradual aspects of cellular fate decisions.
• Automatic Detection and Prediction: Automatically identifies initial, intermediate, and terminal cell populations, computes fate potentials, and visualizes continuous gene expression trends along inferred lineages.
• Application in Diverse Biological Contexts: Applied to datasets such as pancreas development and lung regeneration after injury to detect key populations and predict novel trajectories, including a dedifferentiation trajectory validated experimentally.

Scientific Applications:

• Tissue Regeneration: Maps cell fate decisions during regeneration, for example during lung regeneration after injury, and can reveal dedifferentiation trajectories.
• Disease Progression: Predicts cellular trajectories and fate potentials in disease contexts to inform mechanistic hypotheses.
• Cellular Differentiation and Development: Resolves intermediate states and lineage-specific gene expression trends in developmental systems such as pancreas development.

Methodology:

Calculates RNA velocity from spliced and unspliced mRNA ratios; integrates RNA velocity with trajectory inference; incorporates uncertainty in velocity vectors; computes fate probabilities/potentials and identifies initial, intermediate, and terminal populations while visualizing continuous gene expression trends along lineages.