The reason this second dimension exists is alternative splicing. In mammals, most genes are transcribed as a precursor RNA containing multiple protein-coding segments called exons, and those exons are not always stitched together in the same way. By including some exons, skipping others and occasionally selecting alternative start or end points, a single gene can generate a family of related messenger RNAs — transcript isoforms — that often encode proteins with different functional properties. Neurons are among the body’s heaviest users of this combinatorial freedom: many neural genes exist in dozens of validated isoforms, and isoform choice shapes everything from the proteins that form synapses to the signals that guide developing neurons. Because alternative isoforms can differ in protein domains, localization signals and regulatory sequences, a shift in their relative proportions can change a cell’s behavior without changing how loudly any gene is expressed. The new study centers on a phenomenon called differential transcript usage, or DTU. Unlike differential expression, which asks whether a gene is more or less active in one place than another, DTU asks whether the same overall transcriptional output is being rerouted among a gene’s alternative versions. Conventional spatial transcriptomics, which usually counts short fragments and collapses them to gene-level totals, cannot see that rerouting at all.

Capturing it required solving a three-way trade-off that has defined transcriptomics for a decade. Single-cell RNA sequencing can resolve isoforms in individual cells, but it demands dissociating tissue, which erases every spatial relationship in the cortex, and its short-read chemistries typically capture only fragments near one end of each transcript. Spatial transcriptomics platforms keep molecules anchored in place, but the most widely adopted systems measure compact segments and summarize results at the gene level, leaving isoform identity ambiguous while offering limited fields of view. Long-read sequencers can read an entire transcript from one end to the other in a single pass, resolving isoforms cleanly — but conventional long-read workflows begin from bulk RNA or isolated cells, discarding both positional information and single-cell identity, and their throughput per sample has limited coverage of large tissues. Full-length reads, single-cell resolution and a wide field of view have therefore pulled against one another since the field began, forcing every existing approach to sacrifice at least one. Fullscope-seq was engineered to occupy precisely that gap, and its central innovation is economic rather than exotic: it changes how many full-length molecules a single sequencing read can deliver.

The trick is called programmed concatenation. In a standard long-read workflow, one full-length complementary DNA copy of a transcript is sequenced as one read — an arrangement that becomes costly when the goal is to profile vast numbers of molecules across large expanses of tissue. Fullscope-seq instead threads many full-length cDNA copies, generated from the RNA of a single cell, head to tail into one long concatenated chain, punctuated by short identifying sequences inserted between the copies. Because a modern long-read instrument can read tens of thousands of bases in one pass, a single sequencing read can traverse an entire chain and yield several complete transcripts, which software then separates back into individual full-length molecules. The multiplication substantially increases the number of full-length transcripts captured per run and per unit of tissue, making isoform-resolved mapping of large samples economically feasible. Two additional design choices complete the system. The field of view is large enough to encompass cortical layers and regions within a single measurement, preserving their true anatomical relationships. And the workflow retains the identity of the cell that produced each transcript, so every full-length sequence can be traced back to an individual neuron or glial cell at a measured position. The chemistry was also built, by design, to run on more than one long-read sequencing platform rather than a single instrument.

Applied to the macaque brain, the method produced a kind of data that did not previously exist at this scale: isoform-resolved, spatially addressed, cell-resolved maps of a primate cortex. The headline figure is measured in thousands. Thousands of genes showed differential transcript usage across cortical layers, cell types and brain regions — their relative isoform proportions shifting systematically along multiple anatomical axes even where total gene expression appeared stable. Beyond these gradual redistributions, the team resolved hundreds of major isoform switches, cases in which the predominant transcript of a gene changes outright between one brain region and another, less a change in ratio than a change in identity. The researchers also identified clear differential transcript usage between the superficial layers of the cortex, which receive and integrate incoming signals, and the deep layers, which carry the cortex’s output to distant targets. That vertical span measures only millimeters. Across it, genes alter not merely how abundantly they are transcribed but which of their alternative messages predominate — a form of anatomical patterning that gene-counting surveys compress out of existence.

The maps also address a question that shadows every spatial measurement: when a molecular pattern follows anatomy, does it reflect space itself, or simply which cells happen to occupy that space? The team separated the two explanations. Isoform choices specific to individual cortical layers proved to depend largely on cellular composition — they tracked which cell types populate each layer and could largely be accounted for by that cellular census. Isoform differences between brain regions did not reduce so neatly. Regional differential transcript usage was regulated by both cellular composition and spatial context, meaning that even after accounting for the mix of cell types present, a transcript’s isoform proportions still depended on where in the brain its cell sat. The distinction carries a conceptual consequence: isoform regulation is not solely a property of cell identity. The same cell type, positioned in different spatial settings within the primate cortex, does not necessarily splice its transcripts identically, implicating the local tissue environment — the neighborhood a cell occupies — as an active participant in deciding which version of a gene gets made.

The disease implications sharpen the significance. When the researchers examined the genes carrying spatially patterned isoform variations, the set proved substantially enriched for genes associated with neuropsychiatric disorders. Dysregulated splicing has been moving toward the center of research on conditions such as autism and schizophrenia for years, but nearly all of that evidence comes from tissue that has been dissociated for analysis, stripped of the layered architecture in which cortical computation actually takes place. The new data restore the missing coordinate. They show that the isoform usage of vulnerability-linked genes is not randomly distributed across a healthy primate cortex; it is organized by layer, by cell type and by region. The result does not establish a mechanism, and an isoform pattern in healthy tissue is not a disease process. But it repositions the question. If genes tied to neuropsychiatric illness change their transcript editions along anatomical gradients, then studies that measure genes without locations, or transcripts without full length, risk averaging away precisely the variation most worth studying.

Two further results argue that the patterns reflect biology rather than artifact. When the data were re-examined across different sequencing platforms, the isoform patterns held, ruling out any single instrument’s chemistry as the source of the geography. The patterns also showed conservation across species, indicating that spatially organized transcript choice is not an idiosyncrasy of one animal but a stable feature of complex brains. What the study hands the field is therefore twofold. The first part is a framework: a scalable set of laboratory and computational procedures, from programmed concatenation of full-length cDNA to the software that decomposes concatenated reads back into individual transcripts, which other laboratories can adapt to their own tissues and platforms. The second is a resource — a catalog of full-length, spatially resolved, cell-type-specific transcript usage in the primate brain — that researchers can mine for hypotheses about how cortical layers and regions differ at the level of individual transcript versions. The logic is not brain-specific: any complex tissue in which cell composition and spatial position might jointly shape isoform decisions becomes a candidate for the same analysis.

What emerges is a second cartography of the brain. The first map — which genes are switched on, where, and in which cells — took years to assemble and reorganized neuroscience. The new work overlays a complementary one, recording which version of each transcript occupies which cell at which address, and showing that this record follows rules of its own: some written by the identities of the cells present, others by location itself, and a disproportionate number running through genes tied to neuropsychiatric illness. Until now, the physics of sequencing forced a choice among full-length reads, single-cell identity and broad spatial coverage, and every study gave something up. A protocol that dissolves the trade-off shifts the basic unit of brain genomics from the gene to the specific transcript version inside a specific cell in a specific place. That shift will reach beyond neuroscience, because any complex tissue — developing, aging, diseased — may conceal a comparable isoform geography. The primate cortex is simply the first place anyone has looked closely enough to draw it.

Recent advancements in single-cell and spatial transcriptomics, as well as multi-omics technologies, have propelled the field forward, allowing researchers to probe the intricate molecular landscapes of cells. This surge in understanding is being significantly accelerated by artificial intelligence (AI), which is adept at analyzing vast datasets generated by these high-resolution techniques. As AI continues to evolve, it offers researchers the ability to identify patterns and correlations within complex biological systems that would previously have been imperceptible.

However, the journey toward unlocking the full potential of cellular modeling faces formidable challenges. As researchers strive to capture the dynamic behaviors of cells over time, the computational demands of processing the resulting data often lead to significant bottlenecks. This is where the integration of quantum computing enters the conversation. While still in its relative infancy, quantum computing offers a novel computation paradigm capable of tackling these challenges head-on.

Quantum computing fundamentally differs from classical computing by leveraging the principles of quantum mechanics. This allows quantum computers to perform computations at an exponentially faster rate than their classical counterparts. For applications in biological research, this could mean simulating complex cellular interactions and dynamics with a level of accuracy and speed that is currently unattainable. As quantum technologies advance, we might find ourselves on the brink of a new era in single-cell analysis and modeling.

One particularly exciting aspect of quantum computing is its potential to augment existing AI approaches in the life sciences. By combining classical AI algorithms with the immense computational power of quantum machines, researchers can develop solutions capable of crunching multitudes of parameters more efficiently. This synergy can lead to deeper insights into cellular responses and behaviors under various perturbations, making it an essential tool for precision medicine.

The exploration of how quantum computing can be applied in cell-based therapeutics is already underway. For instance, drug discovery efforts could be revolutionized by utilizing quantum algorithms to model and predict how individual cells or populations respond to specific compounds. This would enable researchers and clinicians to tailor therapeutic interventions to the unique profiles of patients, heralding a new age of personalized medicine that is more effective and responsive.

The cumbersome computational loads often associated with high-resolution biological data can be alleviated with quantum computing, which stands to dramatically reduce the time and resources required for analyses. By enabling faster simulations and data analyses of cellular behaviors, quantum computing could rapidly advance our understanding of disease mechanisms and treatment responses, ultimately leading to improved patient outcomes.

Moreover, the integration of quantum technologies in biomedical research also presents substantial prospects for enhancing collaborations across scientific disciplines. The convergence of biology, physics, and computer science offers a rich tapestry for innovation. Scientists from diverse backgrounds can come together to create hybrid models that utilize the strengths of each field to solve complex biological questions that have long been obstructed by computational limitations.

Despite the tantalizing advantages offered by quantum computing, the journey toward broad adoption will encounter hurdles that the scientific community must address. These include developing quantum algorithms that are specifically tailored to biological applications, ensuring that researchers are equipped with the skills needed to utilize these advanced technologies, and establishing robust frameworks for data sharing and collaboration.

As quantum computing continues to evolve, emerging applications in single-cell analysis will materialize. This will not only impact fundamental research but also influence clinical practices. Biomarker discovery, patient stratification, and therapeutic monitoring are just a few areas where quantum-enhanced analytics could significantly streamline processes.

The potential for quantum computing to refine our understanding of cellular behaviors is immense. In the coming years, we may witness groundbreaking studies that leverage both quantum and classical computing methods alongside high-resolution assays to unlock the intricate tapestry of cellular life. The implications of such advancements extend far beyond the laboratory, fundamentally reshaping the landscape of healthcare.

Ultimately, as we stand on the threshold of this new frontier, it is crucial for the scientific community to engage in ongoing dialogues around the ethics and practical applications of quantum computing in biomedicine. By fostering discussions about the responsible integration of these technologies, researchers can ensure that advancements are used to benefit society as a whole.

In conclusion, the collaboration of high-resolution assays, artificial intelligence, and quantum computing presents an extraordinary opportunity for the field of precision medicine. The ability to generate highly accurate models of cellular dynamics not only promises to enhance our understanding of biology but also holds the potential to revolutionize therapeutic interventions. As we navigate this uncharted territory, the onus is on researchers to harness these advancements responsibly and effectively.

Subject of Research: Quantum Computing in Single-Cell Analysis and Cell-Based Therapeutics

Article Title: Advancing single-cell omics and cell-based therapeutics with quantum computing

Article References:

Bose, A., Rhrissorrakrai, K., Utro, F. et al. Advancing single-cell omics and cell-based therapeutics with quantum computing.
Nat Rev Mol Cell Biol (2026). https://doi.org/10.1038/s41580-025-00918-0

Image Credits: AI Generated

DOI:

Keywords: Quantum Computing, Single-Cell Analysis, Precision Medicine, AI, Multi-Omics, Cell-Based Therapeutics.