RNA Velocity for Predicting Future Cellular States

You are reading version 1 of this direction, frozen 2026-10-04. The latest version is version 2.

Version 1 of 240 papers42 conceptsfrozen 2026-10-04

Standard single-cell RNA sequencing (scRNA-seq) provides a high-resolution "snapshot" of a tissue, capturing the abundance of messenger RNA (mRNA) in thousands of individual cells at a single moment. However, biological processes like development, immune responses, and cancer progression are dynamic "movies." To understand where a cell is going, researchers have developed RNA velocity, a computational framework that treats static transcriptomic data as a directional vector field.

From Snapshots to Vector Fields

The core challenge in single-cell analysis is that the sequencing process is destructive; once a cell is measured, its future is lost. Traditional methods like "pseudotime" attempt to solve this by ordering cells along a path of transcriptional similarity. As Brian Hie, Alex K. Shalek, and colleagues note in Computational Methods for Single-Cell RNA Sequencing, while these paths show potential connections, they lack inherent directionality. We might see a path between a stem cell and a neuron, but the data alone doesn't prove which way the "arrow" of time points.

RNA velocity provides this arrow by looking at the "metabolic" state of mRNA within the cell. By distinguishing between unspliced pre-mRNA (newly transcribed) and spliced mature mRNA (ready for translation), researchers can estimate the derivative of gene expression.

The Kinetic Logic of Splicing

The intuition behind RNA velocity is simple:

  1. Induction: If a gene is being turned "on," the cell will have a high ratio of unspliced to spliced mRNA. The "new" transcripts are being made faster than they can be processed.
  2. Repression: If a gene is being turned "off," the unspliced transcripts will disappear first, leaving a surplus of mature, spliced mRNA that has not yet degraded.

By modeling these transcriptional kinetics, researchers like Xuemei Xie and Qiang Shi can predict the future state of a cell—typically on the scale of hours—from a single measurement. This was famously demonstrated in Single-cell transcriptome profiling reveals neutrophil heterogeneity and orchestrated maturation during homeostasis and bacterial infection, where velocity revealed how infection "overclocks" immune cell production without changing their fundamental identity.

Mapping Developmental Hierarchies

This predictive power allows scientists to reconstruct complex "family trees" of cells. In the skin, Ziguang Lin and Suoqin Jin used RNA velocity to challenge the traditional "stepwise" model of epidermal formation, showing instead a fluid, gradual transition governed by master regulators like GRHL3 (Murine interfollicular epidermal differentiation is gradualistic).

Similarly, in cancer, velocity helps identify the "source" of a tumor. Teresa G. Krieger and colleagues applied these tools to pancreatic cancer organoids, discovering that different clinical subtypes are actually different stages of a shared developmental hierarchy (Single-cell analysis of patient-derived PDAC organoids). By seeing which cells have the "velocity" to move toward more aggressive states, researchers can identify the specific leukemia-initiating cells or oncogenic subpopulations that drive disease (Differentiation hierarchy in adult B cell acute lymphoblastic leukemia at clonal resolution).

Identifying the Origins of Disease

RNA velocity is particularly effective at catching cells in "transitional" states before they exhibit physical symptoms. Tracy Tabib and Minxue Jia have used this to pinpoint the exact moment a healthy fibroblast begins to turn into a scarring myofibroblast in systemic sclerosis and pulmonary fibrosis (Early events marking lung fibroblast transition to profibrotic state in idiopathic pulmonary fibrosis).

In inflammatory diseases like Crohn's, velocity analysis by Jia Li and Alan J. Simmons revealed that specialized "LND" cells don't just appear; they result from the reprogramming of mature cells under stress (A Specialized Epithelial Cell Type Regulating Mucosal Immunity and Driving Human Crohn’s Disease).

New Frontiers: Spatial and Clinical Dynamics

The field is now moving beyond simple splicing ratios. Philipp Weiler and Fabian J. Theis introduced CellRank 2, which unifies velocity with other data types like metabolic labeling to distinguish between different regulatory mechanisms (Unified fate mapping in multiview single-cell data).

Other researchers are integrating physical location into these models. spVelo and SpaceFlow (developed by Qing Nie and others) combine gene expression with spatial proximity to map how cells move through "pseudo-spatiotemporal" maps (Identifying multicellular spatiotemporal organization of cells with SpaceFlow).

Perhaps most strikingly, the concept is moving into the clinic. VeloCD, developed by Aubrey J. Cunnington and colleagues, applies RNA velocity to bulk whole-blood samples to predict the future clinical trajectory of patients with sepsis or COVID-19, turning a single blood draw into a forecast of patient recovery (Predicting trajectories of acute illness using RNA velocity of whole blood).

Papers in this version

History & citationVersion 2 · 2026-10-04

Current · Permanent link

  1. Version 2 (latest) · 2026-10-04 · 32 papers
  2. Version 1 · 2026-10-04 · 40 papers
Cite this direction

Latest version (a living page):

@misc{lacuna_rna_velocity_analysis_750b61b2,
  title = {RNA velocity analysis},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/rna-velocity-analysis/rdr_750b61b2fadc412291fbf8942a02f043},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_rna_velocity_for_predicting_future_cellu_750b61b2_v1,
  title = {RNA Velocity for Predicting Future Cellular States},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/rna-velocity-for-predicting-future-cellular-states/rdr_750b61b2fadc412291fbf8942a02f043/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}