Generalizing RNA velocity to transient cell states through dynamical modeling

Journal: Nature Biotechnology

Published: 2020-08-03

DOI: 10.1038/s41587-020-0591-3

Affiliations: 2

Authors: 5

Go to article

Research Highlight

Forecasting the future state of cells with AI

© MR.Cole_Photographer/Getty

© MR.Cole_Photographer/Getty

AI modelling of gene dynamics in single cells could be used to predict how individual cells will develop.

RNA velocity is the speed at which the expression of each gene in a cell is increasing or decreasing at a given moment. Conventional single-cell sequencing methods measure it only from static snapshots, limiting its use for predicting how a cell will develop.

Now, a team led by researchers at the Technical University of Munich has created an open-access AI model called scVelo to estimate the RNA velocity of dynamic biological systems.

They used machine learning methods to teach scVelo the likelihood of every possible way a gene could act at any point during a cell’s development. The team used scVelo to analyse cell development in the mouse pancreas and in regenerating human lung tissues.

scVelo could be used to study cellular responses to disease progression and cancer treatment.

Supported content

  1. Nature Biotechnology 38, 1408–1414 (2020). doi: 10.1038/s41587-020-0591-3
Institutions Share
Helmholtz Zentrum München - German Research Center for Environmental Health (HMGU), Germany 0.50
Technical University of Munich (TUM), Germany 0.50