AI reveals faulty “off switches” in gene regulation as a key driver of Alzheimer’s disease

Researchers at the National Institutes of Health (NIH) have used artificial intelligence (AI) models of the human genome to uncover a previously underappreciated genetic mechanism that drives Alzheimer’s disease: the failure of “silencers”—regulatory elements that normally keep gene activity in check. The team found that disruption of these silencers leads to gene overactivation in the brain, particularly in the immune cells of patients with dementia, contributing to disease progression.

 While Alzheimer’s disease research has largely focused on coding variants that directly affect genes, including APOE4, over 90% of Alzheimer’s-associated genetic variants lie in the noncoding regions of our DNA. These regions are often referred to as the genome’s “dark matter” because their biological roles have not been fully established.

 Scientists from the NIH’s National Library of Medicine employed an AI-based framework that integrates bulk and single-cell epigenomic data to decode the complex regulatory landscape of brain cells in Alzheimer’s patients. Whereas prevailing models of disease pathology emphasize disrupted gene expression due to coding variants or alterations in enhancers (positive regulatory elements), this study reveals a pronounced role for altered silencers resulting in impaired gene repression.

 The authors identified over 1,000 silencer variants clustered in several regions of the human genome. Genes in these regions are largely inactive in healthy brain tissue but are highly expressed in Alzheimer’s disease, especially in microglia—the brain’s resident immune cells that play a central role in inflammation and immune defense.

 Importantly, the newly developed framework pinpoints likely causal regulatory variants and accurately distinguishes them from nearby benign genomic variants, with results supported by experimental data. This approach opens new paths toward better understanding Alzheimer’s and other dementias, as well as developing therapeutic strategies that target specific variants in the human genome.

 “For years, we’ve known that most Alzheimer’s-linked genetic variants are hidden in the genome’s ‘dark matter,’ but we lacked the tools to see what they were actually doing,” said Dr. Di Huang, lead author. “By using deep learning to integrate single-cell data, we were able to zoom in on these specific genetic ‘brakes’—or silencers—that have failed. It’s a bit like discovering that a car is accelerating out of control not because the gas pedal is stuck, but because the brakes have been cut. This failure of repression allows pro-inflammatory genes in microglia to run wild, driving the chronic neuroinflammation that characterizes the disease.”

 “This work marks a shift in how we prioritize genetic targets for Alzheimer’s disease and dementia in general. By accurately modeling the regulatory code of noncoding DNA, we can distinguish causal variants from the surrounding genetic noise,” said Dr. Ivan Ovcharenko, senior author. “Most importantly, we can characterize the impact of enhancers and silencers simultaneously, allowing us to investigate the mechanisms of aberrant gene expression that lead to Alzheimer’s. This level of precision is exactly what is needed for the next generation of Alzheimer’s therapies targeting specific noncoding variants in the human genome. We can now look toward strategies that directly focus on these broken genomic switches, potentially stopping the overactivation of harmful genes before they cause irreversible damage to the brain.”

About the National Library of Medicine (NLM): NLM is a leader in research in biomedical informatics and data science and the world’s largest biomedical library. NLM conducts and supports research in methods for recording, storing, retrieving, preserving, and communicating health information. It creates resources and tools that are used billions of times each year by millions of people to access and analyze molecular biology, biotechnology, toxicology, environmental health, and health services information. Additional information is available at https://www.nlm.nih.gov.

 About the National Institutes of Health (NIH): NIH, the nation's medical research agency, includes 27 Institutes and Centers and is a component of the U.S. Department of Health and Human Services. NIH is the primary federal agency conducting and supporting basic, clinical, and translational medical research, and is investigating the causes, treatments, and cures for both common and rare diseases. For more information about NIH and its programs, visit www.nih.gov.

 Reference: Di Huang, Ivan Ovcharenko, Silencer variants are key drivers of gene up-regulation in Alzheimer’s disease. Sci. Adv.12, eadz3323 (2026). DOI: 10.1126/sciadv.adz3323