NatureGenomicsPeer-reviewed journal article
RNA velocity in single cells
Gioele La MannoiD, Ruslan SoldatoviD, Amit ZeiseliD, Emelie BrauniD, Hannah HochgerneriD, Viktor PetukhoviD, Katja LidschreiberiD, Maria Eleni KastritiiD, Peter LönnerbergiD, Alessandro FurlaniD, Jean FaniD, Lars E. BormiD, Zehua LiuiD, David van BruggeniD, Jimin GuoiD, Xiaoling HeiD, Roger A. BarkeriD, Erik SundströmiD, Gonçalo Castelo‐BrancoiD, Patrick CrameriD, Igor I. AdameykoiD, Sten LinnarssoniD, Peter V. KharchenkoiD
Abstract
RNA abundance is a powerful indicator of the state of individual cells, but does not directly reveal dynamic processes such as cellular differentiation. Here we show that RNA velocity--the time derivative of RNA abundance--can be estimated by distinguishing unspliced and spliced mRNAs in standard single-cell RNA sequencing protocols. We show that RNA velocity is a vector that predicts the future state of individual cells on a timescale of hours. We validate the accuracy of RNA velocity in the neural crest lineage, demonstrate its use on multiple technical platforms, reconstruct the branching lineage tree of the mouse hippocampus, and measure RNA kinetics in human embryonic brain. We expect RNA velocity to greatly aid the analysis of developmental lineages and cellular dynamics, particularly in humans.
Abstract summary
Claude Opus 5.5 · max effort · 17 Oct 2026 · from the abstract
Unspliced versus spliced mRNA in standard scRNA-seq gives RNA velocity, a per-cell vector predicting future state over hours, validated in neural crest and used to trace mouse hippocampal lineages.
RNA abundance indicates a cell's current state but does not show dynamic processes such as differentiation directly. Estimating RNA velocity, the time derivative of RNA abundance, from the unspliced and spliced mRNA already captured by standard scRNA-seq protocols adds a direction of change to every cell without a new experimental protocol.
Key results
ShownSeparating unspliced from spliced mRNA reads in standard scRNA-seq protocols allows RNA velocity to be estimated as a per-cell vector that predicts the cell's future state on a timescale of hours.
The abstract reports that the accuracy of RNA velocity was validated in the neural crest lineage; no accuracy metric or comparison is given there. · Abstract
ShownApplied to the mouse hippocampus, RNA velocity recovered its branching lineage tree.
Named application in the abstract; no comparison with other lineage inference methods or with independent lineage data is described. · Abstract
ShownThe method worked on data from multiple technical platforms and was used to measure RNA kinetics in human embryonic brain.
Stated as demonstrations in the abstract; the platforms are not named and no kinetic estimates are reported there. · Abstract
Methods, data and reproducibility
Unspliced and spliced mRNA reads from standard scRNA-seq protocols are distinguished to estimate RNA velocity, the time derivative of each cell's RNA abundance, as a vector pointing to its future state; the estimate is validated in the neural crest lineage and applied to mouse hippocampus and human embryonic brain.
- Single-cell RNA-seq data from the neural crest lineage, used for validation
- Single-cell RNA-seq data from mouse hippocampus, used for lineage tree reconstruction · mouse
- Single-cell RNA-seq data from human embryonic brain, used to measure RNA kinetics · human
- Data from multiple technical platforms (not named in the abstract)
- Methods & tools
- RNA velocity, scRNA-seq (standard protocols)
- Reproducibility gaps
- The abstract gives no code, data links or accession numbers.
Limitations (2)
- Accuracy is reported as validated in a single lineage, the neural crest, with no metric given in the abstract. · noted in this summary
- That RNA velocity will greatly help analysis of developmental lineages, especially in humans, is an expectation of the authors rather than a result shown in the abstract. · noted in this summary
Read this if you study differentiation or other dynamic processes with scRNA-seq and want to infer where each cell is heading from the spliced and unspliced reads you already have.
- Citations
- 4,118
- FWCI
- 94.73
- Field percentile
- top 1%
Reproducibility ledger
Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.
Sign in to file a reproducibility report.
Related papers
- Towards reliable quantification of cell state velocities
Valérie Marot-Lassauzaie, Brigitte Joanne Bouman, … Laleh Haghverdi·PLoS Computational BiologyPeer-reviewed·Sep 28, 2022
- A guide to trajectory inference and RNA velocity
Philipp Weiler, Koen Van den Berge, … Simone Tiberi·Methods in molecular biologyPeer-reviewed·Dec 10, 2022
- UniTVelo: temporally unified RNA velocity reinforces single-cell trajectory inference
Mingze Gao, Chen Qiao, Yuanhua Huang·Nature CommunicationsPeer-reviewed·Nov 3, 2022
- DeepVelo: Deep Learning extends RNA velocity to multi-lineage systems with cell-specific kinetics
Haotian Cui, Hassaan Maan, … Bo Wang·Genome biologyPeer-reviewed·Jan 19, 2024
- Pyro-Velocity: Probabilistic RNA Velocity inference from single-cell data
Qian Qin, Eli Bingham, … Luca Pinello·bioRxiv·Sep 12, 2022
- Improving the RNA velocity approach using long-read single cell sequencing
Chen Zhang, Weitian Chen, … Chong Tang·bioRxiv·May 2, 2022