8
0RegVelo is an advanced deep learning framework designed to study the complex developmental dynamics of single cells by integrating gene regulatory networks (GRNs) with RNA velocity. Unlike previous models that viewed gene transcription in isolation, this method treats cellular transitions as a regulated process, allowing for a more accurate reconstruction of cell fate decisions. By combining scRNA-seq data with prior biological knowledge, the tool creates a predictive, in silico "cell" capable of simulating how specific genetic perturbations influence development. Researchers utilized this framework to identify and validate key lineage drivers in biological systems like zebrafish neural crest and human hematopoiesis. Ultimately, RegVelo bridges the gap between regulatory circuitry and physical cell transitions, offering a quantitative way to generate and test mechanistic hypotheses in systems biology.
References:
Wang W, Hu Z, Weiler P, et al. RegVelo: gene-regulatory-informed dynamics of single cells[J]. Cell, 2024.
