What's new and why it matters:
- A first-of-its-kind brain atlas. Researchers mapped 232,251 individual cells across four stages of development, revealing both which genes are active and the DNA switches that control them. Earlier developmental atlases of the fly brain captured the genes, but not that regulatory layer.
- A neuron's identity is established during a brief window just after it is born. Rather than simply inheriting a finished program from their parent cell, newly born neurons rapidly open genetic switches that establish their lifelong identity
- There is no single "master switch." Neuronal identity emerges from a specific combination of regulators and enhancers, explaining how the nervous system can generate many cell types from a limited genetic toolkit.
- The findings could help researchers repair, replace, or even design neurons. By revealing more about the "recipe" that produces specific types of neurons, the study could inform scientists on how to guide cells toward the neuron needed in various diseases, including Parkinson's, ALS, and glaucoma.
KANSAS CITY, Mo. — September 10, 2026 — A brain contains hundreds or thousands of different kinds of neurons, each with its own shape, connections, and function. Yet all of them arise from a comparatively small pool of dividing neural stem cells. How does a newly born cell know what kind of neuron to become?
A new study from the Stowers Institute for Medical Reseearch and the lab of Assistant Investigator Neşet Özel, Ph.D. , published in Proceedings of the National Academy of Sciences on August 19, 2026, is revealing the genetic logic behind that decision. Özel's lab created the most detailed developmental atlas yet of the fruit fly (Drosophila) visual system, following 232,251 individual cells across the course of development.
"This atlas is, as all atlases are, simply a map," Özel said. "It shows what genes are expressed by each neuron, and what genomic regions, or switches, controlling that expression are open in each neuron. Sometimes two cells express the same gene, but not through the same switch. This tells us something new about development and cell fate."
That last point lead to an unexpected discovery. It revealed a popular concept of development — one that has shaped neuroscience for decades — is likely more nuanced than scientists long understood it to be.
"We discovered the switchboard logic that gives each kind of neuron its lasting identity," Özel said. "What was especially striking is that this identity is established during a brief critical period right after a neuron is born — when the regulatory landscape of its DNA is dramatically remodeled — rather than simply inherited from its parent cell."
The findings revealed that many enhancers associated with neuronal identity were completely closed in the parent stem cell and opened only after the neuron completed its final cell division. In other words, the neuron is not simply carrying forward a finished identity program from its parent cell.
"We wanted to understand one of the oldest puzzles in neuroscience: how a brain builds hundreds of different kinds of neurons, each with its own shape, wiring, and function, starting from a small pool of stem cells," Özel said. "In particular, how does a newly born cell 'decide' which type of neuron to become, and how is that decision written in the DNA?"
The answer, the researchers found, is not a single genetic command. Instead, neuronal identity is built through a flexible, modular system in which combinations of regulatory proteins act through different DNA switches in different neurons.
Such a distinction has implications beyond developmental biology: If scientists want to generate or replace a specific type of neuron, they need to understand not just which genes define that neuron, but how and when those genes are activated.
"It's very clear that medicine needs to make specific types of neurons to treat various diseases, and to make a specific type of neuron, you need to know that neuron's recipe," Özel said. "That recipe, as it turns out, is not that easy to come up with. We are trying to understand how it is written in the genome, and how you could use that information to make the cells medicine needs most."
Identity is computed, not inherited
The atlas revealed a missing transition step in neuronal development and one that Özel refers to as "the most striking finding of the paper." The team showed that identity is not simply passed down from a stem cell but, instead, actively rebuilt after the neuron is born.
"The DNA landscape is dramatically remodeled in the first hours of a neuron's life," Özel said. "That tells us identity isn't simply copied from the stem cell. In an important sense, it is computed during a specific critical period."
Scientists have long known that combinations of proteins, called transcription factors, help establish neuronal identity. Some of these factors, known as terminal selectors, maintain the characteristics that distinguish one neuron type from another throughout the life of the cell.
Özel had previously mapped many of these identity codes across the development of the fruit fly's visual system. What remained unclear was how those codes were installed in the first place.
The new study shows that even when the same regulatory protein appears in both a stem cell and one of its neuronal descendants, it may be activated through entirely different DNA switches before and after the cell's final division.
"That same protein would be doing something else entirely in the progenitor, and the genome will not respond to it in the same way it responds to it in the neuron," Özel said. "Right now, this is not really clear in the field. It's not widely appreciated."
No master switch
The scale of the atlas also allowed the researchers to ask whether different neuron types share a common regulatory logic.
They found much less uniformity than expected.
"When we inferred the gene regulatory networks for different groups of neurons, we expected to find more consistency in terms of which factors control which genes across the brain," said McKenzie Treese, first author of the study and a Computational Biology Scholar in the Özel Lab. "Instead, almost every neuron type had its own wiring."
The same transcription factor could regulate largely different sets of genes from one neuron type to another. Conversely, the same gene could respond to different transcription factors through different enhancers depending on the cell.
Rather than a single master regulator dictating identity, each neuron appears to use its own combination of regulatory inputs.
"The regulation is deeply context dependent," Özel said. "The same gene can be controlled by different switches in different neurons, and each neuron type uses its own combination of regulators. That flexibility gives the nervous system a way to generate enormous cellular diversity from a limited genetic toolkit."
The researchers also found that enhancer use changes dramatically over development and is highly specific to individual cell types, so specific that the collection of accessible switches in a cell can function almost like a fingerprint of neuronal identity.
Learning how to build neurons
Many neurological conditions such as Parkinson's disease, ALS, and glaucoma involve the loss of particular types of neurons. Replacing those cells requires understanding how to make the right neuron, not any neuron.
"Before therapy, we need to understand the fundamentals," Özel said. "That requires decades of research and building up principles."
The concept extends to neurodevelopmental disorders, many of which trace back to these same regulatory proteins and DNA switches going awry during the very narrow developmental window.
"It took this specific technology, looking at two different genomic modalities at once, exactly during this transition, for us to gain these insights," Özel explained. "We need a deeper understanding of how one builds a brain to solve all these problems. This provides us the map and the resource to go after that 'how' question."
Related research : A companion study from the Özel Lab builds on findings from this work and tests one of the atlas's predictions in greater detail. The study is available as a preprint on bioRxiv and has not yet been peer reviewed. Read the preprint.
A marble rolling down the hill : Explore an interactive breakout that explains Özel's finding, offering a new understanding and nuance to a long-standing view of development. View here.
Additional authors include Yen-Chung Chen; Abigail Tyree; Rose Coyne; Cathleen Lake; Ojong Besong Tabi; Raghuvanshi Rajesh; Yu-Chieh David Chen; Huzaifa Hassan; Hua Li; and Claude Desplan, Ph.D.
This work was funded by the National Institute of Neurological Disorders and Stroke of the National Institutes of Health; the National Eye Institute of the National Institutes of Health; and the Esther A. and Joseph Klingenstein Fund.
About the Stowers Institute for Medical Research
Founded in 1994 through the generosity of Jim Stowers, founder of American Century Investments, and his wife, Virginia, the Stowers Institute for Medical Research is a non-profit, biomedical research organization with a focus on foundational research. Its mission is to expand our understanding of the secrets of life and improve life's quality through innovative approaches to the causes, treatment, and prevention of diseases.
The Institute consists of 25 independent research programs. Of the approximately 500 members, over 370 are scientific staff that include principal investigators, fellows, technology center directors, postdoctoral scientists, graduate students, and technical support staff. Learn more about the Institute at www.stowers.org and about its graduate program at www.stowers.org/gradschool .
