CellRank for directed single-cell fate mapping.

Area of Science:

• Single-cell genomics
• Systems biology
• Computational biology

Background:

• Trajectory inference methods reconstruct cell state dynamics from single-cell RNA sequencing (scRNA-seq).
• Current methods require predefined directionality, limiting applications to normal developmental processes.
• Inferring cell fate in complex scenarios like regeneration, reprogramming, or disease remains challenging.

Purpose of the Study:

• To introduce CellRank, a computational tool for single-cell fate mapping applicable to diverse biological systems, even when the direction of the process is unknown.
• To integrate RNA velocity with trajectory inference to capture stochastic and gradual cell fate decisions.
• To provide a robust framework for analyzing cell state dynamics across various experimental contexts.

Main Methods:

• CellRank combines established trajectory inference techniques with directional information derived from RNA velocity.
• The method accounts for the inherent stochasticity and gradual nature of cellular fate transitions.
• Uncertainty in RNA velocity vectors is explicitly modeled to improve robustness.

Main Results:

• CellRank successfully identified initial, intermediate, and terminal cell populations in pancreas development data.
• The tool accurately predicted cell fate potentials and visualized continuous gene expression dynamics along inferred lineages.
• Application to lineage-traced reprogramming data validated CellRank's ability to predict reprogramming outcomes.
• CellRank uncovered a novel dedifferentiation trajectory in lung regeneration, including previously unidentified intermediate cell states, which were experimentally confirmed.

Conclusions:

• CellRank provides a powerful and versatile approach for single-cell fate mapping across various biological contexts, including those with unknown directional information.
• The integration of RNA velocity enhances the accuracy and scope of trajectory inference, enabling the study of complex cellular dynamics.
• CellRank facilitates the discovery of novel cellular states and trajectories, advancing our understanding of developmental, regenerative, and disease processes.