Abstract
Rich data from large biobanks, coupled with increasingly accessible association statistics from genome-wide association studies (GWAS), provide great opportunities to dissect the complex relationships among human traits and diseases. We introduce BADGERS, a powerful method to perform polygenic score-based biobank-wide association scans. Compared to traditional approaches, BADGERS uses GWAS summary statistics as input and does not require multiple traits to be measured in the same cohort. We applied BADGERS to two independent datasets for late-onset Alzheimer’s disease (AD; N=61,212). Among 1,738 traits in the UK biobank, we identified 48 significant associations for AD. Family history, high cholesterol, and numerous traits related to intelligence and education showed strong and independent associations with AD. Further, we identified 41 significant associations for a variety of AD endophenotypes. While family history and high cholesterol were strongly associated with AD subgroups and pathologies, only intelligence and education-related traits predicted pre-clinical cognitive phenotypes. These results provide novel insights into the distinct biological processes underlying various risk factors for AD.
Introduction
Identifying the factors that contribute to late-onset Alzheimer’s Disease (AD) is a critical challenge in modern medicine. Traditional epidemiological studies, while valuable, are often hindered by high costs and the risk of "reverse causality"—where it is unclear if a lifestyle factor causes the disease or if early-stage disease causes changes in lifestyle.
The emergence of massive biobanks, such as the UK Biobank, and large-scale Genome-Wide Association Studies (GWAS) offers a potential solution. However, a "data mismatch" often occurs: a large GWAS for Alzheimer’s may not include the same individuals who were measured for the thousands of other traits (e.g., diet, blood markers, lifestyle) found in a biobank.
The BADGERS Framework
To address this, researchers developed BADGERS (Biobank-wide Association Discovery using GEnetic Risk Scores). This computational framework allows for "hypothesis-free" screening of thousands of potential risk factors by using genetics as a proxy.
BADGERS operates as a two-stage "genetically-imputed" association method:
- Imputation: Using Biobank data, the method generates Polygenic Risk Scores (PRS) for thousands of traits, creating a genetic "fingerprint" for each.
- Association: These genetic fingerprints are then tested for association with the disease of interest (in this case, AD) using only GWAS summary statistics.
A key advantage of BADGERS is its efficiency. Unlike traditional methods that require restricted individual-level genotype data, BADGERS works entirely with summary statistics and an external reference panel (like the 1000 Genomes Project) to account for Linkage Disequilibrium.
Key Findings: Risk Factors and Endophenotypes
The researchers applied BADGERS to two independent AD datasets (IGAP and ADGC) and 1,738 heritable traits from the UK Biobank. The scan identified 48 significant associations for AD, including family history, high cholesterol, and various cognitive and educational markers.
The study also explored "endophenotypes"—specific biological markers of the disease. They identified 41 significant associations linked to cerebrospinal fluid (CSF) biomarkers and neuropathological scores, such as Braak stages.
The "Two-Path" Model of AD
A significant insight from this work is the divergence in how different risk factors impact the disease:
- Physical Pathology: Factors like family history and high cholesterol were strongly linked to actual brain pathology (amyloid plaques and tau tangles). These associations were heavily influenced by the APOE gene, the strongest genetic risk factor for AD.
- Cognitive Resilience: In contrast, intelligence and education-related traits did not necessarily correlate with physical brain changes. Instead, they predicted better cognitive performance (e.g., in the WRAP pre-clinical cohort) despite the presence of the disease. This supports the "cognitive reserve" hypothesis, suggesting these traits act as a buffer.
Conclusion and Future Directions
BADGERS proved to be a superior "first-pass" tool, identifying significantly more candidate risk factors than traditional Mendelian Randomization (MR) methods. While the method identifies associations rather than direct causality, it provides a powerful way to prioritize traits for further investigation.
Future work will focus on applying BADGERS to more diverse ancestries and expanding the library of traits as biobank data continues to grow. By bridging the gap between massive genetic datasets and clinical traits, BADGERS offers a new lens through which to understand the complex biology of Alzheimer’s Disease.
Summary updated