Summary: Researchers have created a new computational tool that provides direct genetic evidence linking Alzheimer’s disease to the loss of memory-forming neurons. The algorithm, called seismic, maps genetic risk to individual cell types more precisely than earlier methods, clarifying long-standing contradictions in dementia research by showing that specific neurons — not only immune cells — are implicated in Alzheimer’s pathology.
By integrating large-scale genetic studies with single-cell expression data, the tool offers a sharper view of how inherited risk factors translate into vulnerability of particular cells. The team behind seismic says the approach can be applied broadly to other neurological and metabolic diseases, improving detection of disease-relevant cell populations and informing therapeutic strategies.
Key Facts:
- New algorithm: Seismic combines genome-wide association data with single-cell RNA expression to identify which cell types drive disease risk.
- Clear genetic link: The method reveals genetic evidence that ties Alzheimer’s risk to memory-related neurons rather than primarily to immune cells.
- Wide applicability: The technique is designed to work across many complex disorders, including Parkinson’s disease, to help guide research and drug development.
Source: Rice University
The global burden of dementia was estimated at 57 million people in 2021, with about 10 million new cases each year. In the United States, over 6 million people live with dementia, and new cases are projected to rise substantially in the coming decades, according to a 2025 study.
Despite progress in diagnosis and care, the molecular and cellular mechanisms that cause dementia remain incompletely understood.
To close this knowledge gap, researchers at Rice University, in collaboration with Boston University, developed a computational framework to determine which specific cell types are genetically linked to complex human traits and diseases, including Alzheimer’s and Parkinson’s. The framework, named “Single-cell Expression Integration System for Mapping genetically implicated Cell types” — abbreviated as seismic — pinpoints genetic vulnerabilities in the neurons responsible for forming memories. This is the first study to show a direct genetic association between Alzheimer’s and these specific neuronal populations.
Published in Nature Communications, the study helps reconcile a long-standing contradiction: genetic studies have repeatedly implicated infection-fighting immune cells of the brain (microglia), while neuropathology shows that neurons involved in memory die in Alzheimer’s. The new analysis shows how both observations can be part of the same disease picture by linking genetic risk to particular neuronal populations.
“Aging naturally slows some brain functions, but dementia involves the irreversible loss of specific neurons,” said Qiliang Lai, a Rice doctoral student and the study’s first author. “Our findings clarify why DNA-based signals and post-mortem brain observations seemed to point in different directions.”
The team’s method combines two complementary data types: genome-wide association studies (GWAS), which detect genetic variants associated with disease risk across large populations, and single-cell RNA sequencing (scRNA-seq), which profiles gene activity at single-cell resolution across tens of thousands to millions of individual cells. By integrating these datasets, seismic maps how polygenic risk converges on discrete cell types and brain regions.
Earlier integrative attempts faced two major limitations. First, single-cell datasets can be analyzed at overly broad cell-type resolutions that obscure region-specific or subtype-specific signals. Second, GWAS based on clinical diagnoses often emphasize cell types that show the most consistent downstream changes, such as immune cells, potentially masking other disease-relevant cell types. Seismic addresses these issues by computing a specificity score that captures both expression magnitude and consistency across cell types, and by performing influential gene analysis to identify the genes driving each cell type–trait association.
Benchmarking showed that seismic outperformed existing tools, producing clearer and more interpretable associations between genetic risk and specific cell populations. In Alzheimer’s and Parkinson’s analyses, the framework revealed cell- and brain-region-specific patterns of vulnerability and identified molecular pathways implicated in neuronal degeneration.
“This approach could help reconcile contradictory patterns in Alzheimer’s genetics and pathology,” said Vicky Yao, assistant professor of computer science at Rice and a member of the Ken Kennedy Institute. “More broadly, it offers a scalable, interpretable way to discover which cell types matter most for many complex diseases.”
The work arrives alongside growing public investment in brain health research. In Texas, the state legislature established the Dementia Prevention and Research Institute of Texas (DPRIT) to accelerate innovation in prevention, treatment, and care. A proposed statewide funding measure would invest significantly in that effort over the coming decade.
“Advances in computing and data science are reshaping disease research,” Yao added. “Sustaining that momentum will be essential to translate discoveries into prevention and treatment.”
Funding: This research received support from the National Institutes of Health (RF1AG054564, R21AG085464), CPRIT (RR190065), the Cure Alzheimer’s Fund and the Karen Toffler Charitable Trust. The content reflects the authors’ views and not necessarily those of the funders.
Key Questions Answered:
A: The team developed seismic, a computational method that links polygenic risk from GWAS to specific cell types using single-cell expression data, enabling precise identification of cells implicated by genetic risk in diseases like Alzheimer’s.
A: The method reconciles a discrepancy between genetic studies that implicate immune cells and neuropathological evidence of neuronal loss by showing a direct genetic association between Alzheimer’s risk and memory-forming neurons.
A: Seismic integrates GWAS and single-cell RNA-seq data, computing a specificity score for cell-type expression and identifying influential genes that drive the association between genetic risk and particular cell types and brain regions.
A: By revealing which cell types are most affected by genetic risk, the approach can guide targeted research into disease mechanisms, improve biomarker discovery, and focus therapeutic development on the most relevant cellular targets.
About this AI, genetics, and Alzheimer’s disease research news
Author: Silvia Cernea Clark
Source: Rice University
Contact: Silvia Cernea Clark – Rice University
Image: The image is credited to Neuroscience News
Original Research: Open access. “Disentangling associations between complex traits and cell types with seismic” by Qiliang Lai et al., Nature Communications
Abstract
Disentangling associations between complex traits and cell types with seismic
Combining single-cell RNA sequencing with genome-wide association studies can reveal which cell types participate in complex traits and disease. Existing methods, however, often struggle with scalability, interpretability, and robustness.
We introduce seismic, a framework that defines a specificity score capturing both expression magnitude and consistency across cell types, and an influential gene analysis to pinpoint genes driving each cell type–trait association.
Across more than 1,000 cell-type characterizations and 28 polygenic traits, seismic confirms known links and uncovers trait-relevant cell groups missed by other approaches.
In Parkinson’s and Alzheimer’s disease analyses, seismic highlights cell- and brain-region-specific differences in pathology and, when applied to pathology-based Alzheimer’s GWAS, identifies vulnerable neuronal populations and molecular pathways involved in neurodegeneration.
Overall, seismic offers a computationally efficient, interpretable, and powerful approach for mapping relationships between polygenic traits and cell-type-specific expression, delivering new insights into disease mechanisms.