Related Experiment Video
Updated: Mar 27, 2026
10:55
Live Imaging Followed by Single Cell Tracking to Monitor Cell Biology and the Lineage Progression of Multiple Neural Populations
Published on: December 16, 2017
9.2K
Interpretable learning of temporal cellular dynamics from single-cell data
Idris Kouadri Boudjelthia1, Salvatore Milite2, Nour El Kazwini3
1Theoretical and Scientific Data Science, International School for Advanced Studies, Trieste, Italy; Abdus Salam International Centre for Theoretical Physics, Trieste, Italy.
Cell Reports Methods
|March 24, 2026
Summary
NeuroVelo reconstructs cell fate dynamics using neural ordinary differential equations and RNA velocity. This method accurately predicts cell transitions and identifies gene regulatory networks driving cell fate.
Area of Science:
- Computational Biology
- Genomics
- Systems Biology
Background:
- Reconstructing cellular dynamics from static single-cell transcriptomics is challenging.
- Current RNA velocity methods have limitations in interpretation and predictive power.
Purpose of the Study:
- To develop a novel method, NeuroVelo, for improved reconstruction of temporal cellular dynamics.
- To simultaneously determine cell transitions and identify gene interactions driving these dynamics.
Main Methods:
- NeuroVelo couples optimal linear projection with non-linear neural ordinary differential equations.
- It utilizes dynamical systems theory in an optimized latent space.
Main Results:
- NeuroVelo accurately reconstructs cell-type transitions from single-cell data.
- The method successfully identifies gene regulatory networks driving cell fate.
- Benchmarking shows NeuroVelo outperforms existing state-of-the-art methods.
Conclusions:
- NeuroVelo offers a powerful approach for understanding temporal cellular dynamics.
- It provides a unified framework for inferring cell fate trajectories and regulatory mechanisms.
