The study, led by Amanda A. G. Ferreira, Sergio Córdoba, Raghuvanshi Rajesh, Ben Jiwon Choi, and Claude Desplan, drew on two large single-cell mRNA-sequencing datasets covering the developing Drosophila visual system. Single-cell sequencing allows scientists to read out the complete transcriptional profile of thousands of individual cells, effectively giving every cell a molecular identity card. Rather than grinding up tissue and averaging signals across cell types, the technique preserves the distinctions between neighboring cells, making it possible to spot rare populations and to follow how one cell type transforms into another across developmental time. The team comprehensively annotated the glial cells within these atlases, identifying every glial cell type present from larval stages through to the adult fly.

With that catalog in hand, the researchers could reconstruct developmental trajectories, the branching paths by which immature glia mature into specialized subtypes. The picture that emerged was not uniform. Most glial types, such as the chiasm glia that occupy the crossings where axon bundles reroute between layers of the optic lobe, follow a gradual trajectory: their transcriptomes shift steadily as they mature, with no dramatic discontinuity between larval and adult states. This smooth maturation suggests that these cells refine an existing identity over time, tuning their gene expression as the circuits around them assemble and begin to function.

Neuropil glia, however, told a very different story. In the larval optic lobe, these cells appear transcriptionally as a single, coherent class, indistinguishable in the data from one another. But as the animal enters pupal stages, the period of dramatic metamorphosis when the fly’s nervous system is extensively remodeled, this single class splits into two distinct types: ensheathing glia, which wrap around and compartmentalize neural processes, and astrocyte-like glia, which resemble the star-shaped astrocytes of the mammalian brain and are thought to interact closely with synapses. The finding provides a concrete example of how glial diversity is actively generated during development rather than simply inherited from a correspondingly diverse set of progenitors.

Crucially, the team did not rely on computational predictions alone. They experimentally validated the developmental trajectories they inferred from the sequencing data and identified genetic markers that are expressed along the differentiation path separating ensheathing glia from astrocyte-like glia. Such markers are valuable currency in developmental biology: they allow other researchers to specifically label, isolate, or manipulate each glial subtype, turning a descriptive atlas into a functional toolkit. For the Drosophila community, which has long used the fly visual system as a model for wiring specificity, neural development, and glia-neuron interactions, these markers open the door to targeted studies of what each glial type actually does as circuits form.

The most unexpected discovery came from a second look at the sequencing data itself. Standard single-cell RNA sequencing is usually interpreted under the assumption that each cell yields one transcriptome. But glial cells are architecturally unusual: they possess a compact cell body from which extend elaborate processes, thin branches that can travel considerable distances to contact neurons. When the researchers analyzed their datasets with this geometry in mind, they found that the transcriptome of the glial cell body could be distinguished from that of its processes. In other words, the messenger RNA content of a single glial cell is not uniformly distributed but is spatially partitioned between its structural compartments.

This kind of subcellular mRNA compartmentalization has important implications for how glia operate. Localizing specific transcripts to processes means that the machinery for producing particular proteins can be positioned close to where those proteins are needed, allowing a cell to respond locally to signals from neurons without waiting for instructions from the cell body. Neurons are well known to use such local translation in dendrites and axons, but demonstrating a comparable system in glia, and doing so at the level of a whole developmental atlas, adds a new dimension to how scientists think about glial autonomy. Glia may not simply react to neuronal cues; they may be equipped to make spatially precise decisions within their own branches.

To make this analysis rigorous, the researchers developed an innovative computational approach designed to identify mRNA species that are differentially localized to cell bodies or to cellular processes. The method addresses a genuine technical challenge: dissociating tissue for single-cell sequencing can shear off processes, so distinguishing true biological compartmentalization from technical artifacts requires careful modeling. The team then validated their computational predictions in vivo, confirming in living flies that the mRNAs flagged by their approach really do accumulate in the compartments the algorithm predicted. That combination of a novel analytical tool and direct experimental confirmation strengthens the claim that subcellular RNA sorting is a real and widespread feature of optic lobe glia rather than a quirk of the dataset.

The work, published in PLOS Biology, represents the most detailed transcriptomic analysis of optic lobe glia conducted during development, and it helps explain a puzzle that has intrigued researchers studying the adult fly brain: why the adult visual system contains such an expanded repertoire of glial types compared with earlier stages. The answer, according to the atlas, lies in processes like the pupal splitting of neuropil glia into ensheathing and astrocyte-like classes, which multiply glial diversity as the animal matures. Development, in this view, is not merely a matter of growth but of progressive diversification, with glial lineages branching into increasingly specialized roles as the circuits they support take shape.

Beyond what it says about flies, the study offers a template that could travel. The logic of combining large single-cell atlases, trajectory inference, and compartment-aware transcriptomic analysis applies to any tissue with architecturally complex cells, including the glia of the mammalian brain, where astrocytes and oligodendrocytes likewise extend far-reaching processes. And because many fundamental mechanisms of neural development are conserved between insects and vertebrates, the molecular markers and developmental principles uncovered here may guide hypotheses about how glial diversity arises in the human nervous system as well. What began as an exercise in cataloging fly glia has ended up revealing that even the brain’s support staff are organized with a precision that rivals the neurons they serve, down to the individual RNA molecules they sort into their farthest-reaching branches.

Subject of Research: Developmental single-cell transcriptomics of Drosophila optic lobe glia

Article Title: A developmental single-cell atlas of the Drosophila visual system glia reveals cell type diversification and subcellular mRNA compartmentalization

Article References: Ferreira, A. A. G., Córdoba, S., Rajesh, R., Choi, B. J., & Desplan, C. (2026). A developmental single-cell atlas of the Drosophila visual system glia reveals cell type diversification and subcellular mRNA compartmentalization. PLOS Biology, 24(9), e3004015. https://doi.org/10.1371/journal.pbio.3004015

Image Credits: AI Generated

DOI: 10.1371/journal.pbio.3004015

Keywords: Drosophila, glia, single-cell RNA sequencing, optic lobe, developmental biology, ensheathing glia, astrocyte-like glia, mRNA localization, cell fate, neurodevelopment, PLOS Biology, transcriptome atlas