Disease doesn't break one thing—it cascades through an interconnected system.
For generations, medicine has studied disease organ by organ, as if the body were a collection of separate rooms rather than a single dwelling. Now, a team at Helmholtz Munich has built MouseMapper, an AI system capable of reading an entire transparent mouse body at cellular resolution, revealing how obesity quietly dismantles nerve architecture and immune order across every tissue at once. The work, published in Nature, marks a shift from the partial view to the whole — and in doing so, suggests that what we have called isolated diseases may be ripples in a single, interconnected disruption.
- Obesity's damage to nerves and immune cells spans the entire body simultaneously, yet science has lacked any tool capable of seeing that full picture until now.
- MouseMapper combines fluorescent cell labeling, tissue-clearing that renders mice transparent, and deep learning to automatically map 31 organ types, nerve pathways, and immune clusters across a whole organism in three dimensions.
- Obese mice showed dramatically fewer branches in the trigeminal nerve — a major facial nerve — and responded less to sensory stimulation, forging a direct, measurable link between structural nerve damage and lost feeling.
- Critically, the same molecular signatures of nerve inflammation and remodeling found in obese mice appeared in trigeminal tissue from obese humans, raising the possibility that this damage crosses species.
- The team has released their full-body datasets publicly, and the lead researcher envisions building cellular-level digital twins of mice in health and disease — turning the body into a navigable, manipulable map.
Scientists have long understood that obesity does far more than alter a person's shape — it rewires immune function, warps nerve structure, and scrambles tissue organization in ways that open pathways to diabetes, heart disease, stroke, and cancer. The obstacle was never knowledge of the damage, but the inability to see it whole. Researchers could examine individual organs, but not an entire living organism at once, at the cellular level, in three dimensions.
That constraint has now shifted. Ali Ertürk's team at Helmholtz Munich built MouseMapper, an AI system trained on a foundational model that generalizes beyond its original data. The pipeline works in stages: nerves and immune cells are marked with fluorescent dyes, the animals are rendered transparent through tissue-clearing techniques, and light-sheet microscopy produces detailed 3D images of the entire body. MouseMapper then analyzes those images automatically — identifying 31 organ types, tracing nerve paths, and locating immune cell clusters without human guidance.
When the system was turned on mice fed a high-fat diet, it revealed widespread disruption in immune organization and nerve architecture across the body. The most striking finding involved the trigeminal nerve, which governs facial sensation and motor control. Obese mice had far fewer nerve branches than lean ones — and behaved accordingly, responding less to sensory stimulation. The imaging and the lived experience of the animals aligned.
The researchers then looked deeper, using spatial proteomics to examine the molecular environment of the trigeminal ganglion. The signatures of nerve remodeling and inflammation they found there appeared in trigeminal tissue from obese humans as well — a bridge between mouse model and human disease.
The team has released their datasets publicly for global collaboration, and Ertürk's ambition reaches further: he wants to construct cellular-level digital twins of mice in health and disease, atlases that scientists could explore and manipulate entirely in silico. MouseMapper is the first instrument that allows researchers to stop studying disease as a failure in one room and begin seeing it as a disturbance moving through the whole house.
Scientists have long known that obesity damages far more than a person's waistline. It rewires immune function, warps nerve structure, and scrambles tissue organization across the entire body—pathways that lead to diabetes, heart disease, stroke, nerve damage, and cancer. Yet for years, researchers lacked the tools to see these changes in full detail across a living organism without destroying it in the process. They could study individual organs or tissues, but not the whole animal at once, at the cellular level, in three dimensions.
That constraint has just shifted. A team led by Ali Ertürk at the Helmholtz Munich Institute for Biological Intelligence and Ludwig Maximilian University has built MouseMapper, an artificial intelligence system that can analyze biological imaging data from an entire mouse body and automatically map what it finds. The system identifies 31 different organs and tissue types, traces the paths of nerves, and locates immune cells scattered throughout the organism—all without human researchers having to tell it where to look.
The technical feat required several pieces to fit together. First, the team marked the nerves and immune cells in mice with fluorescent dyes visible under a microscope. Then they applied tissue-clearing techniques that rendered the animals transparent while preserving the fluorescent signals. Using advanced light-sheet microscopy, they generated detailed three-dimensional images of entire mice. The resulting datasets contained tens of millions of cellular structures across different organs and tissues. MouseMapper then analyzed these images automatically, identifying nerves, immune cell clusters, and anatomical regions throughout the body. The system works because it was trained on a foundational model—meaning it generalizes far beyond the specific data it learned from, as first author Ying Chen explained in the study published in Nature.
To study obesity's effects, the researchers fed mice a high-fat diet that produced obesity and metabolic dysfunction similar to what occurs in humans. MouseMapper revealed widespread disruption in how immune cells were organized and how nerve structures were arranged across the body. One finding stood out: the trigeminal nerve, a major facial nerve responsible for sensation and motor control, showed far fewer branches and endings in obese mice than in lean ones. This structural damage correlated with behavior—obese mice responded less to sensory stimulation than their lean counterparts, establishing a direct link between what the imaging showed and what the animals could actually feel.
The researchers then examined the trigeminal ganglion, where the cell bodies of facial sensory neurons live. Using spatial proteomics, they identified molecular changes associated with nerve remodeling and inflammation. Strikingly, many of these same molecular signatures appeared in trigeminal tissue from obese humans, suggesting that the nerve damage observed in mice might occur in people too.
The implications extend well beyond obesity. MouseMapper could transform how scientists study any complex disease affecting multiple organ systems—diabetes, cancer, neurodegenerative conditions, autoimmune disorders. Unlike previous methods that focused on selected tissues or organs, MouseMapper offers an integrated platform for analyzing the whole body at once, capable of finding disease hotspots anywhere in the organism. The team has released the full-body datasets publicly online, inviting researchers worldwide to study how obesity alters tissues and organ systems.
Ertürk's vision reaches further still. He wants to build realistic digital twins of mice in health and disease—cellular-level atlases that researchers could examine, manipulate, and track entirely in silico. The goal is to understand disease not as isolated failures in individual organs, but as disruptions rippling through an interconnected system. MouseMapper is the first tool that lets scientists actually see the whole picture at once.
Notable Quotes
MouseMapper generalizes far beyond the data it was initially trained on because it uses a foundational model— Ying Chen, first author of the study
The goal is to create a comprehensive framework that lets us understand how diseases affect the body as an interconnected system, eventually building realistic digital twins of mice in health and disease— Ali Ertürk, director of the Helmholtz Munich Institute for Biological Intelligence
The Hearth Conversation Another angle on the story
Why does it matter that we can see the whole mouse at once instead of studying pieces separately?
Because obesity doesn't just break one thing. It cascades. Your immune cells migrate, your nerves shrivel, your tissues reorganize. If you only look at the pancreas, you miss what's happening in the fat or the liver or the nerves. You need the whole map to understand the pattern.
But couldn't researchers just study each organ individually and add up the findings?
Not really. The interactions matter. A nerve doesn't exist in isolation—it's embedded in tissue, surrounded by immune cells, bathed in inflammatory signals. You can't understand what's happening to it without seeing its context. That's what MouseMapper gives you.
The fact that they found the same molecular signatures in human tissue—how confident should we be that this applies to people?
It's a strong signal but not proof. Mice aren't humans. But when you see the same damage pattern and the same molecular changes in both, it suggests the mechanism is real. It's enough to make human studies worth doing.
What does it mean that they made the mice transparent?
It's the only way to image deep structures without slicing the animal into sections. You lose the three-dimensional relationships that way. Transparency lets you see how everything connects in space, all at once.
Why release the data publicly?
Because the real work is just beginning. Thousands of researchers can now ask their own questions about obesity without having to generate these massive datasets themselves. It accelerates discovery across the field.