Overview
- Researchers introduced RegVelo in Cell as a deep-learning framework that models how cells change over time and which gene interactions drive those shifts.
- RegVelo integrates RNA velocity, which reads a cell’s direction from unspliced versus spliced RNA, with gene regulatory networks to map trajectories and forecast the impact of switching specific regulators off.
- In zebrafish neural crest development, the model highlighted tfec as an early pigment driver and pointed to elf1 as a new regulator, findings confirmed with CRISPR/Cas9 knockouts and single-cell Perturb-seq.
- The work draws on an experimental–computational partnership across the Stowers Institute, Helmholtz Munich, TUM, and the University of Oxford, with first author Weixu Wang leading the framework’s design.
- The team cites uses in cancer, developmental disorders, and regenerative medicine, while noting simplifying assumptions, significant compute demands, and the need for broader, multimodal validation.