The way DNA is folded inside certain brain cells differs in people with Alzheimer's disease, according to a study published in Science by researchers at Carnegie Mellon University's School of Computer Science, the University of Pittsburgh School of Medicine and the University of Washington. The team tied those differences in genome folding to shifts in which genes were switched on and to how the brain tissue itself was organized.
Genome folding is an often-overlooked part of how cells work. The same DNA sequence sits in every cell, but the physical loops and compartments it forms determine which stretches come into contact with the switches that turn genes on and off. Finding that this architecture is disturbed in Alzheimer's brain cells points to a regulatory layer sitting between genetic risk and the disease process that has not been mapped in much detail.
Reaching that level of resolution required stitching together three approaches: single-cell technology to look at individual cells rather than averaged tissue, spatial mapping to keep track of where those cells sit in the brain, and a newly built deep learning model to make sense of the combined data. Jian Ma, the Ray and Stephanie Lane Professor of Computational Biology, who led the work, said Alzheimer's cannot be understood one layer at a time, and described the genome's 3D structure as a fundamental regulatory layer connecting DNA sequence to gene activity.
The practical hope is that mapping how these layers fit together narrows the field of candidate mechanisms worth testing, rather than simply adding more entries to the long catalog of changes observed in Alzheimer's brains. As Ma framed it, the goal is to move from listing disease-associated differences to working out which mechanisms deserve experimental follow-up.