Researchers at UC Irvine have built the first causal gene regulatory maps for Alzheimer’s disease, using an AI system called SIGNET to identify which genes actually drive harmful changes in the brain - rather than merely appearing alongside them.
The findings, published in Alzheimer’s & Dementia, analyzed brain samples from 272 participants and revealed extensive rewiring of genetic interactions, particularly in excitatory neurons. The work points toward specific “hub genes” that could become targets for earlier diagnosis and future therapies.
Beyond Correlation to Causation
Most genetic analysis tools detect genes that move together statistically. SIGNET (Statistical Inference on Gene Regulatory Networks) does something harder: it identifies true cause-and-effect relationships while accounting for feedback loops.
“Unlike many traditional tools that simply flag gene associations, SIGNET identifies directional relationships,” said Min Zhang, who led the research with Dabao Zhang at UC Irvine’s Joe C. Wen School of Population & Public Health. The distinction matters for drug development - you want to target genes that cause problems, not genes that merely react to them.
The team integrated single-cell RNA sequencing with whole-genome sequencing data from the Religious Orders Study and Rush Memory and Aging Project, a long-running study of older adults.
What They Found
SIGNET constructed causal regulatory networks across six major brain cell types: excitatory neurons, inhibitory neurons, astrocytes, microglia, oligodendrocytes, and oligodendrocyte progenitor cells.
The most dramatic disruptions appeared in excitatory neurons, where approximately 6,000 genetic interactions showed extensive rewiring as the disease progresses. This suggests excitatory neurons experience the greatest regulatory upheaval in Alzheimer’s.
The researchers identified hundreds of “hub genes” that function as central controllers, affecting many downstream genes. Because of their broad influence, these hubs may play an outsized role in driving disease-related damage. Notably, the APP gene - already known to be involved in Alzheimer’s - demonstrated “strong control” of other genes in inhibitory neurons, validating the method’s ability to flag biologically relevant targets.
Results were confirmed using an independent set of human brain samples, strengthening confidence in the observed mechanisms.
What This Means
Most Alzheimer’s research has focused on the amyloid and tau proteins that accumulate in diseased brains. SIGNET’s approach looks upstream - at the genetic control centers that may be driving those accumulations and other cellular dysfunction.
If specific hub genes are identified as causal drivers in different brain cell types, they become candidates for:
- Early biomarkers (detectable before symptoms appear)
- Drug targets (disrupting the cascade before damage spreads)
- Patient stratification (different genetic subtypes might need different treatments)
The UC Irvine team says SIGNET can be applied beyond Alzheimer’s to other complex diseases, including cancer, autoimmune disorders, and mental health conditions.
The Fine Print
This is discovery-phase research. Identifying a causal hub gene doesn’t mean we know how to safely modify it, or that doing so would reverse disease progression. Brain gene regulation is layered and redundant - blocking one hub might shift the burden elsewhere.
The study analyzed post-mortem brain samples, which captures end-stage disease better than early stages. The team validated findings in independent samples, but replication in larger and more diverse populations will be needed.
The National Institute on Aging and National Cancer Institute funded the research.
Reference: Zhang M, Zhang D, et al. Causal gene regulatory networks in Alzheimer’s disease across cell types. Alzheimer’s & Dementia. 2026;22(2). DOI: 10.1002/alz.71053