Scientists have built some spectacularly powerful tools for interrogating genes - tools that can knock out, dial up, or eavesdrop on thousands of genes at once. CRISPR screens and massively parallel reporter assays (MPRAs) have been tearing through the mysteries of gene regulation like undergraduates through free pizza. There's just one small hitch: these tools were designed for cells growing in a dish. The brain, that 1.4-kilogram lump of opinionated tissue between your ears, has other plans.

A new review in Trends in Neurosciences by Din Selmanovic and Joseph Dougherty at Washington University lays out precisely why the brain keeps rejecting our best molecular interrogation techniques - and, rather encouragingly, how researchers are starting to outwit it (Selmanovic & Dougherty, 2026).

The Greatest Hits That Don't Translate

Here's the setup. In cultured cells, you can deliver a library of tens of thousands of CRISPR guide RNAs, knock out genes wholesale, then read out the consequences at single-cell resolution using something called Perturb-seq. It's like sending a questionnaire to every cell in the dish and getting back nicely filled-in answers. Massively parallel reporter assays do something similar for gene regulation: they test thousands of DNA sequences simultaneously to see which ones switch genes on and where.

In a dish, this is straightforward. Cells are uniform, plentiful, and cooperative. The brain is none of these things.

Why Your Neurons Are Terrible Lab Partners

The review identifies two central headaches, and they're substantial.

Problem one: getting the goods in and the data out. To deliver genetic libraries into the brain, researchers rely on viral vectors - typically adeno-associated viruses (AAVs). But AAVs have a strict carry-on luggage policy. Their packaging capacity maxes out around 4.7 kilobases, which is uncomfortably tight when you need to fit in a CRISPR guide, a reporter, a barcode, and all the regulatory bits that make the thing actually work. On top of that, viral delivery to the brain is sparse: you can't just dunk a brain in virus and expect 90% infection like you can with cultured cells. Signal recovery is correspondingly thin. You're essentially trying to read a library book through a keyhole.

Problem two: the brain is absurdly complex. A cubic millimetre of cortex contains dozens of neuronal subtypes, each with distinct gene expression programmes, layered into circuits that respond dynamically to activity. Testing a genetic element in "brain cells" is a bit like testing a medication on "Europeans" - you've glossed over rather a lot of meaningful variation. Cell-type specificity, spatial organisation, and activity-dependent gene regulation all demand a level of resolution that bulk assays simply cannot provide.

The Clever Workarounds

The good news is that neuroscientists are a stubborn lot, and they've been making progress on both fronts.

For the delivery problem, new AAV-based platforms like CrAAVe-seq have achieved cell-type-specific CRISPR screening at unprecedented scale, targeting over 5,000 genes in mouse brains by cleverly exploiting Cre-sensitive guide RNA constructs (Nature Neuroscience, 2025). Meanwhile, in vivo Perturb-seq has leapt from labelling less than 0.1% of brain cells with lentiviruses to over 6% using optimised AAV serotypes and transposon systems - a 60-fold improvement that actually makes large-scale screening feasible (Cell, 2024; PMID: 38772369).

For the complexity problem, researchers are layering CRISPR screens onto single-cell and spatial transcriptomics. Spatial Perturb-seq now allows scientists to knock out genes while preserving tissue architecture, revealing not just what a gene does to the cell it's in, but how that ripples through neighbouring cells (Nature Communications, 2026). On the enhancer side, Cre-dependent MPRAs enable cell-type-specific readouts of regulatory elements in living brains, a far cry from the one-size-fits-all approach of earlier assays (Communications Biology, 2023).

Why You Should Care (Even If You Never Touch a Pipette)

The non-coding genome - the 98% of DNA that doesn't encode proteins - is where most disease-associated genetic variants actually live. These variants don't break proteins; they fiddle with the volume knobs that control when, where, and how much protein gets made. Psychiatric disorders, neurodegenerative diseases, and neurodevelopmental conditions are overwhelmingly linked to these regulatory regions, and we cannot properly decode them without testing their function in the actual tissue that matters.

Every advance in making high-throughput assays work in living brains brings us closer to understanding why certain genetic variants raise the risk of schizophrenia, Alzheimer's, or autism - not in a dish, but in the organ where these conditions actually unfold.

The brain has always been the most reluctant organ to give up its secrets. It appears, however, that persistence and some rather ingenious molecular engineering are beginning to wear down its defences.

References

1. Selmanovic, D., & Dougherty, J. D. (2026). Challenges and opportunities in adapting high-throughput functional assays for in vivo neuroscience. Trends in Neurosciences, 49(3), 217-232. DOI: 10.1016/j.tins.2026.01.008 | PubMed

2. CRISPR screening by AAV episome-sequencing (CrAAVe-seq): a scalable cell-type-specific in vivo platform uncovers neuronal essential genes. (2025). Nature Neuroscience. DOI: 10.1038/s41593-025-02043-9

3. Massively parallel in vivo Perturb-seq reveals cell-type-specific transcriptional networks in cortical development. (2024). Cell. PMID: 38772369. PMC: PMC11193654

4. Spatial perturb-seq: single-cell functional genomics within intact tissue architecture. (2026). Nature Communications. DOI: 10.1038/s41467-026-69677-6

5. A Cre-dependent massively parallel reporter assay allows for cell-type specific assessment of the functional effects of non-coding elements in vivo. (2023). Communications Biology. DOI: 10.1038/s42003-023-05483-w

6. Assembloid CRISPR screens reveal impact of disease genes in human neurodevelopment. (2023). Nature. DOI: 10.1038/s41586-023-06564-w

7. CRISPR-Cas9 screens reveal regulators of ageing in neural stem cells. (2024). Nature. DOI: 10.1038/s41586-024-07972-2

Disclaimer: The image accompanying this article is for illustrative purposes only and does not depict actual experimental results, data, or biological mechanisms.