In a preprint, researchers built CAMBUS, a machine-learning method that models how DNA changes may affect whether genes can turn on in a cell. Across 29 immune cell types, it found 66,043 regions especially sensitive to sequence disruption, often overlapping DNA that evolution appears to preserve.
This can help researchers focus on non-coding DNA changes that may alter gene activity in particular cells. CAMBUS connected 1,533 possible causal variants across 70 complex traits to cell-specific regulatory regions, including links involving immune cells. The connections identify candidates for study, not proven causes.
✦ Sequence-sensitive DNA regions can help researchers narrow the search for disease mechanisms.