CMU-Pitt Study Links Genome Folding to Alzheimer’s

Researchers at Carnegie Mellon University, the University of Pittsburgh and the University of Washington have linked 3D genome folding to gene activity and brain tissue organisation in Alzheimer’s disease, identifying a previously underexplored regulatory layer of the disease’s biology. Published in Science on 23 July 2026, the multiomics work introduces a new deep learning model, Hicformer, that can predict gene activity from genome architecture in specific brain cell types.

The team analysed postmortem tissue from the prefrontal cortex of individuals with and without Alzheimer’s, drawn from a long-term dementia study. They used GAGE-seq, which measures gene expression and 3D genome contacts in the same cell, integrating results with spatial transcriptomic maps of intact tissue. Hicformer, the group’s new deep learning model, combines DNA sequence, broad genome-folding features and local 3D contact maps to predict gene activity across cell types, functioning as a computational test bed for how altered genome folding may change gene activity.

Large active and inactive regions of the genome, known as compartments, were less clearly separated in cells from people with Alzheimer’s, a pattern the team calls “increased compartment mingling” and one associated with lower overall gene activity. Multiple brain cell types showed fewer short-range contacts and more long-range contacts, while contacts between genes and nearby regulatory elements weakened. These architectural changes were linked to reduced neuronal and synaptic programs, altered metabolic and stress responses, and senescence-related programs in microglia. Hansruedi Mathys of the University of Pittsburgh framed the findings as establishing higher-order chromatin alterations as a component of Alzheimer’s molecular pathology, sitting alongside amyloid-beta plaques and tau tangles.

The commercial signal is a broadening of the Alzheimer’s target landscape beyond amyloid-beta and tau at exactly the moment approved therapies (Eisai and Biogen’s Leqembi, Eli Lilly’s Kisunla, Roche’s Trontinemab) are still delivering modest clinical impact. Chromatin architecture as a druggable layer opens new opportunities for epigenetic-focused biotechs, single-cell multiomics platforms and computational biology teams to enter the field.

For the latest updates and in-depth insights into the world of Life Science, including breakthrough treatments, industry trends, and regulatory news, contact Bailey Osborne today!