A study published in Science has found that Alzheimer’s 3D genome organisation is consistently disrupted across several types of brain cells, adding a layer to the disease’s biology that sits alongside the amyloid plaques and tau tangles that have long dominated research. The findings come from a collaboration involving Carnegie Mellon University’s School of Computer Science, the University of Pittsburgh School of Medicine, the University of Washington, and several other institutions.
What Alzheimer’s 3D genome organisation research found
DNA does not sit in a cell as a straight strand. It folds into a complex three-dimensional structure that governs which genes are accessible and active. The team examined postmortem samples from the prefrontal cortex, taken from people with and without Alzheimer’s disease who had participated in a long-term dementia study and later donated their brains for research. Using a technique called GAGE-seq, which can measure both gene expression and three-dimensional genome contacts within the same individual cell, the researchers then combined those measurements with spatial transcriptomic maps that preserve information about where gene activity occurs within intact tissue.
Several consistent structural differences emerged in cells from people with Alzheimer’s disease. Large sections of the genome are normally organised into relatively distinct active and inactive regions, known as compartments. In Alzheimer’s cells, those boundaries appeared less sharply defined, a pattern the researchers describe as ‘increased compartment mingling’. Brain cells showing greater compartment mingling tended to have lower overall levels of gene activity. The team also observed weaker interactions between genes and nearby regulatory elements that normally help determine whether genes are switched on or off, alongside stronger contacts across intermediate distances.
These structural changes were associated with reduced activity in programmes involved in neurons and synapses, along with alterations in metabolism and cellular stress responses. The researchers also found links to senescence-related programmes in microglia, immune cells in the brain that help maintain brain health and respond to damage.
Chromatin alterations and the disease’s molecular landscape
The study’s claim is that changes in chromatin, the material made of DNA and associated proteins that packages the genome inside cells, should be considered part of Alzheimer’s molecular pathology, not just a secondary effect. ‘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,’ said Hansruedi Mathys, assistant professor of neurobiology at Pitt’s Department of Neurobiology, who directed the Pitt arm of the study.
That is a meaningful distinction. Identifying a new structural layer does not, on its own, establish causation. The researchers are careful about this: they describe Alzheimer’s 3D genome organisation differences as ‘associated with’ changes in gene activity, and the paper frames future work around testing whether particular structural changes help drive disease progression or could eventually become targets for new therapies.
The AI model built to connect structure and function
A separate strand of the research involved an artificial intelligence model called Hicformer, developed to investigate how genome structure may influence cellular behaviour. The model combines DNA sequence information with broad patterns of genome folding and detailed maps of physical contacts between different sections of DNA, then uses those inputs to predict gene activity across different cell types. Xinyue Lu, a doctoral student in Computational Biology who co-led the research, described it as a computational test bed for exploring how changes in genome folding might alter gene activity.
‘Measuring gene activity and genome folding in the same cell allows us to directly connect chromosome structure with disease-related gene programmes,’ said Yang Zhang, a project scientist in the Computational Biology Department who co-led the research. ‘Across several kinds of brain cells, this paired view revealed a consistent signature of 3D genome reorganisation in Alzheimer’s disease and helped us prioritise regulatory regions for future mechanistic and therapeutic investigation.’
The team behind the work
The study was led and supervised by Jian Ma, the Ray and Stephanie Lane Professor of Computational Biology at Carnegie Mellon. According to Carnegie Mellon’s School of Computer Science, Ma received the Allen Newell Award for Research Excellence in 2025. His framing of the findings reflects a broader methodological argument: ‘Alzheimer’s disease cannot be understood one layer at a time,’ he said. ‘By integrating genome folding, cell state, and tissue context, we can move beyond cataloguing disease-associated changes toward understanding how they fit together and which mechanisms to test next.’
The research was supported by grants from the National Institutes of Health. Other Carnegie Mellon authors included doctoral students Shahul Alam and Shike Wang, and postdoctoral research associate Junjie Tang. Other Pitt authors include doctoral students Alexander K. Kunisky and Jude Baroudi, post-baccalaureate research fellows Sahar and Sahel Ghorbanikalateh, and visiting scholar Shihan Wang. The team also included researchers from the Broad Institute of MIT and Harvard, the University of California Los Angeles, the University of Washington, and the Rush Alzheimer’s Disease Center.
The paper is published in Science as ‘Single-cell multiomics connects 3D genome and transcriptome alterations in Alzheimer’s disease’, with DOI 10.1126/science.adz1652. The next step, the researchers say, is to test whether specific structural changes in the genome directly contribute to disease progression.
