Overview
- The paper, published in Science in July 2026, used postmortem prefrontal cortex from donated brains to compare 3D genome folding and gene activity in people with and without Alzheimer's disease.
- Researchers combined GAGE-seq, which measures gene expression and physical genome contacts in the same single cell, with spatial transcriptomics to place folding changes inside intact tissue neighborhoods.
- A new AI model called Hicformer integrated DNA sequence with multi-scale folding features to predict cell-type gene activity and help prioritize regulatory regions for follow-up tests.
- Across multiple brain cell types the team found 'increased compartment mingling'—active and inactive genome regions were less separated—fewer short-range contacts, more long-range contacts, and weakened local gene–regulatory interactions linked to lower neuronal and synaptic programs and stress or senescence signatures in microglia.
- Funded by the NIH and produced by a multi-institution team, the study stops short of proving causation but signals a new layer of Alzheimer’s biology and points to experiments that could reveal structural regulatory targets with potential therapeutic value and impact on how scientists study neurodegeneration.