DNA is not a flat list of instructions. Inside a cell, it is packed into a dense three-dimensional structure, and which genes a cell can actually read depends partly on how that structure is arranged. A study published in Science reports that this architecture is organized differently in several types of brain cells from people who had Alzheimer's disease, and that the differences track with changes in gene activity.

The research came from Carnegie Mellon University's School of Computer Science, the University of Pittsburgh School of Medicine and the University of Washington, with collaborators at the Broad Institute of MIT and Harvard, the University of California, Los Angeles, and the Rush Alzheimer's Disease Center. It was published on July 23 and resurfaced in science coverage this week.

For the roughly seven million Americans living with Alzheimer's and the families caring for them, the honest framing matters. This is mechanistic research. It identifies a layer of biology that had been underexplored and gives researchers a way to prioritize which regions of the genome to test next. It does not change diagnosis, treatment, or care today.

Beyond Plaques and Tangles

Most people who follow Alzheimer's research know two storylines: amyloid-beta plaques that accumulate between neurons, and tau tangles that form inside them. Both remain central, and both have driven drug development programs for the past two decades.

Hansruedi Mathys, an assistant professor of neurobiology at Pitt who directed the university's arm of the study, positioned the new findings as an addition rather than a replacement. "We know the classic hallmarks of Alzheimer's disease, accumulation of amyloid-beta plaques and tau tangles, but our results establish higher-order chromatin alterations as a component of the molecular pathology associated with the disease, which currently affects seven million Americans, a number that continues to grow," Mathys said in the University of Pittsburgh's summary of the study findings.

Chromatin is the combination of DNA and the proteins that package it. The argument is that changes in how that packaging is organized deserve a place alongside plaques and tangles in the description of what goes wrong.

Jian Ma, the Ray and Stephanie Lane Professor of Computational Biology at Carnegie Mellon, who led and supervised the study, described the reasoning in the university's announcement of the work. "Alzheimer's disease cannot be understood one layer at a time," Ma said. "The genome's 3D structure is a fundamental regulatory layer that helps to connect DNA sequence to gene activity. By integrating genome folding, cell state and tissue context, we can move beyond cataloging disease-associated changes toward understanding how they fit together and which mechanisms to test next."

Measuring Folding and Activity in the Same Cell

The study examined postmortem tissue from the prefrontal cortex, the region at the front of the brain, taken from people who had participated in a long-term dementia study and donated their brains to research. According to the National Institutes of Health, which funded part of the work, the team analyzed tissue from 20 people aged 75 or older, 10 of whom had Alzheimer's and 10 of whom did not.

The technical advance is that the team measured two things at once in individual cells. Using a method called GAGE-seq, researchers captured both gene expression and three-dimensional genome contacts in the same cell, then combined those measurements with spatial transcriptomic maps that preserve information about where gene activity occurs within intact tissue. That combination allowed them to connect physical genome organization to gene regulation while keeping track of the surrounding tissue context.

They also built an artificial intelligence model called Hicformer, which combines DNA sequence information with broad genome folding features and local contact maps to predict gene activity across cell types. Xinyue Lu, a computational biology doctoral student who co-led the research, described it as a computational test bed for asking how altered genome folding may change gene activity.

Yang Zhang, a project scientist in Carnegie Mellon's Computational Biology Department who also co-led the work, explained what the paired measurement bought them. "Measuring gene activity and genome folding in the same cell allows us to directly connect chromosome structure with disease-related gene programs," Zhang said. "Across several kinds of brain cells, this paired view revealed a consistent signature of 3D genome reorganization in Alzheimer's disease and helped us prioritize regulatory regions for future mechanistic and therapeutic investigation." Zhijun Duan, a research associate professor, co-led the research from the University of Washington.

Boundaries That Blur

The specific differences have a consistent character. Large stretches of the genome are normally sorted into relatively distinct active and inactive zones called compartments. In cells from people with Alzheimer's, those boundaries were less clearly separated, a pattern the researchers call increased compartment mingling.

Multiple brain cell types also showed fewer short-range contacts and more long-range contacts. Greater compartment mingling was associated with lower overall gene activity. Contacts between genes and the nearby regulatory elements that help switch them on or off were weaker, while some midrange contacts strengthened.

These structural differences were linked to reduced neuronal and synaptic gene programs, altered metabolic and stress responses, and senescence-related programs in microglia, the brain's resident immune cells. When the changes were mapped across intact tissue, genome reorganization tracked not only with altered gene activity but also with differences in how brain cells were organized within the tissue.

Reading the Evidence Honestly

Several limitations belong up front rather than buried. This is an observational study of postmortem brain tissue, which means it captures the end state of a disease that develops over decades. It cannot establish whether the architectural changes contribute to Alzheimer's, result from it, or reflect both. The researchers describe their own contribution as providing a framework for future experiments to determine which changes in genome structure contribute directly to the disease.

The sample is small, 10 donors with Alzheimer's and 10 without, and the study focused on a single brain region, the prefrontal cortex. Other work has found that epigenomic changes in Alzheimer's vary considerably between regions, so findings from the prefrontal cortex may not generalize. The Hicformer predictions are model outputs and require experimental validation. The research was supported by grants from the National Institutes of Health, with no industry funding disclosed.

For anyone hoping this points toward a therapy, the distance is substantial. Identifying regulatory regions worth investigating is several steps removed from a drug target, and many steps removed from a treatment.

Families looking for guidance they can act on should rely on established sources rather than mechanistic research. The National Institute on Aging maintains current information on Alzheimer's and dementia, including diagnosis, care planning and clinical trial participation. Anyone concerned about memory changes in themselves or a relative should raise it with a physician, who can evaluate whether a specialist referral is appropriate. No finding in this study changes that advice.

What Readers Want to Know

What did researchers find? The three-dimensional folding of DNA is organized differently in several types of brain cells from people with Alzheimer's, and those differences track with changes in gene activity.

How is this different from amyloid and tau? Plaques and tangles are protein accumulations. This finding concerns chromatin, the physical packaging of DNA. Researchers present it as an additional layer, not a replacement for the existing hallmarks.

How many people were studied? Postmortem prefrontal cortex tissue from 20 donors aged 75 or older, 10 with Alzheimer's and 10 without, according to the National Institutes of Health.

Does this lead to a treatment? Not directly and not soon. The study identifies regulatory regions worth investigating, which is early-stage work well upstream of any therapy.

What are the main limitations? It is observational postmortem research with a small sample covering one brain region. It cannot show whether the changes cause disease or result from it, and some results come from a computational model needing validation.

Where and when was it published? In Science, on July 23, 2026, by teams at Carnegie Mellon, the University of Pittsburgh, and the University of Washington, with several collaborating institutions.

What should families do with this information? Nothing changes in care or treatment. For guidance, the National Institute on Aging maintains current dementia information, and memory concerns should be raised with a physician.