A European interdisciplinary research project, coordinated by Thibault Lagache (BioImage Analysis Unit, Institut Pasteur – France-BioImaging’s platform), David Menassa (University of Oxford), and David Holcman (University of Cambridge), has recently led to the creation of an innovative tool: DeepCellMap. This tool significantly improves the mapping of microglia, brain cells that remain widely unknown to the general public.

Microglia belong to the family of glial cells, which form the environment around neurons. They provide immune protection for the nervous system and play a crucial role in brain development.

What does DeepCellMap bring?

• The tool can classify microglia into five categories based on their shape and location in the brain. This classification allows researchers to track their roles throughout brain development.
• DeepCellMap also revealed a dynamic spatial organization of microglia. These cells occupy distinct territories and reorganize as the brain grows. Clusters of cells appear and later disperse, following specific patterns.
• An unexpected finding: DeepCellMap uncovered a strong association between microglia and blood vessels in the cerebral cortex of fetuses exposed to SARS-CoV-2 during pregnancy. This discovery raises a major question: Are microglia reacting to vascular alterations, or do they themselves contribute to these changes? This observation could pave the way for new research into the impact of prenatal infections on the developing brain.

How does DeepCellMap work?

This tool relies on a deep learning algorithm (artificial intelligence) capable of detecting and classifying microglia based on their morphology, using bright-field or fluorescence microscopy images.

The analysis of microglial spatial organization is made possible by the use of advanced statistical models.
In conclusion, DeepCellMap is a groundbreaking image analysis tool for biology. By enabling large-scale studies of brain cells, it opens new research perspectives in neuroscience to better understand the mechanisms of brain development.

In the long term, DeepCellMap may also help improve our understanding of how prenatal infections affect the developing brain. As an open-source resource, it can be adapted to study other cell types and applied to a wide range of human health research.

Congratulations to all the teams involved! France-BioImaging is proud to support interdisciplinary projects such as the development of DeepCellMap by members of our bioanalysis community. Supporting research through cutting-edge R&D and fostering international collaboration are at the heart of our infrastructure’s missions.

Want to know more? Click here to read the article: https://www.nature.com/articles/s41467-025-56560-z