Challenges at the APOE locus: a robust quality control approach for accurate APOE genotyping

Alzheimer's Disease Sequencing Project release 4 whole genome sequencing dataset

INTRODUCTION

The Alzheimer's Disease Sequencing Project (ADSP) is a national initiative to understand the genetic architecture of Alzheimer's disease and related dementias (ADRD) by integrating whole genome sequencing (WGS) with other genetic, phenotypic, and harmonized datasets from diverse populations.

METHODS

The Genome Center for Alzheimer's Disease (GCAD) uniformly processed WGS from 36,361 ADSP samples, including 35,014 genetically unique participants of which 45% are from non-European ancestry, across 17 cohorts in 14 countries in this fourth release (R4).

RESULTS

This sequencing effort identified 387 million bi-allelic variants, 42 million short insertions/deletions, and 6.8 million structural variants. Annotations and quality control data are available for all variants and samples. Additionally, detailed phenotypes from 15,927 participants across 10 domains are also provided. A linkage disequilibrium panel was created using unrelated AD cases and controls.

DISCUSSION

Researchers can access and analyze the genetic data via the National Institute on Aging Genetics of Alzheimer's Disease Data Storage Site (NIAGADS) Data Sharing Service, the VariXam, or NIAGADS GenomicsDB.

HIGHLIGHTS

We detailed the genetic architecture and quality of the Alzheimer's Disease Sequencing Project release 4 whole genome sequences. We identified 435 million single nucleotide polymorphisms, insertions and deletions, and structural variants from diverse genomes. We harmonized extensive phenotypes, linkage disequilibrium reference panel on subset of samples. Data is publicly available at NIAGADS Data Storage Site, variants and annotations are browsable on two different websites.

A Quantitative Trait Locus for Reduced Microglial APOE Expression Associates with Reduced Cerebral Amyloid Angiopathy

The Apolipoprotein E (APOE) e4 and e2 alleles are respectively the most risk increasing and risk decreasing, common genetic risk factors for Alzheimer's disease (AD)1,2. They strongly affect Aβ burden in the brain parenchyma1, a core hallmark of AD, but also at the level of the brain vasculature, i.e. cerebral amyloid angiopathy (CAA)1,3, which in turn relates to increased risk for amyloid-related imaging abnormalities (ARIA) in APOE*4 carriers when receiving anti-Aβ antibody treatments4. This makes APOE a highly pursued AD drug target. A crucial question in the field is whether it would be beneficial to either increase or decrease APOE (particularly APOE*4) levels5. The answer from rodent work appears to converge on "decreasing APOE levels"5-7, with initial human studies supporting this5,8,9. Human genetic evidence however remains scarce and new insights are crucially needed to support clinical translation. Shade et al. 2024 conducted the largest to date genome-wide association study (GWAS) of various neuropathological traits, identifying a variant protective of CAA in the APOE locus independent of APOE*4 and APOE*2 genotypes10. Downstream analyses suggested this signal links to the nearby APOC2 gene through local effects on methylation. We applaud the authors on their timely, relevant, and well-conducted study. Here, we extend on these findings, highlighting there is compelling evidence that their genetic signal for reduced CAA relates to an effect on reduced microglial APOE expression, which would importantly support the evidence in favor of "decreasing APOE levels" and further herald this promising therapeutic avenue, not just for AD, but also for CAA. We additionally provide complimentary results regarding this locus' association with CAA and AD risk from analyses that we conducted parallel to Shade et al. 2024.

Beyond guilty by association at scale: searching for causal variants on the basis of genome-wide summary statistics

Understanding the causal genetic architecture of complex phenotypes will fuel future research into disease mechanisms and potential therapies. Here, we illustrate the power of a novel framework: it detects, starting from summary statistics, and across the entire genome, sets of variants that carry non-redundant information on the phenotypes and are therefore more likely to be causal in a biological sense. The approach, implemented in open-source software, is also computationally efficient, requiring less than 15 minutes on a single CPU to perform genome-wide analysis. Through extensive genome-wide simulation studies, we show that the method can substantially outperform existing methods in false discovery rate control, statistical power and various fine-mapping criteria. In applications to a meta-analysis of ten large-scale genetic studies of Alzheimer's disease (AD), we identified 82 loci associated with AD, including 37 additional loci missed by conventional GWAS pipeline. Massively parallel reporter assays and CRISPR-Cas9 experiments have confirmed the functionality of the putative causal variants our method points to. Finally, we retrospectively analyzed summary statistics from 67 large-scale GWAS for a variety of phenotypes. Results reveal the method's capacity to robustly discover additional loci for polygenic traits and pinpoint potential causal variants underpinning each locus beyond conventional GWAS pipeline, contributing to a deeper understanding of complex genetic architectures in post-GWAS analyses.

Alzheimer's Disease Sequencing Project Release 4 Whole Genome Sequencing Dataset

The Alzheimer's Disease Sequencing Project (ADSP) is a national initiative to understand the genetic architecture of Alzheimer's Disease and Related Dementias (AD/ADRD) by sequencing whole genomes of affected participants and age-matched cognitive controls from diverse populations. The Genome Center for Alzheimer's Disease (GCAD) processed whole-genome sequencing data from 36,361 ADSP participants, including 35,014 genetically unique participants of which 45% are from non-European ancestry, across 17 cohorts in 14 countries in this fourth release (R4). This sequencing effort identified 387 million bi-allelic variants, 42 million short insertions/deletions, and 2.2 million structural variants. Annotations and quality control data are available for all variants and samples. Additionally, detailed phenotypes from 15,927 participants across 10 domains are also provided. A linkage disequilibrium panel was created using unrelated AD cases and controls. Researchers can access and analyze the genetic data via NIAGADS Data Sharing Service, the VariXam tool, or NIAGADS GenomicsDB.

Role of the X Chromosome in Alzheimer Disease Genetics

Importance

The X chromosome has remained enigmatic in Alzheimer disease (AD), yet it makes up 5% of the genome and carries a high proportion of genes expressed in the brain, making it particularly appealing as a potential source of unexplored genetic variation in AD.

Objectives

To perform the first large-scale X chromosome-wide association study (XWAS) of AD.

Design, Setting, and Participants

This was a meta-analysis of genetic association studies in case-control, family-based, population-based, and longitudinal AD-related cohorts from the US Alzheimer's Disease Genetics Consortium, the Alzheimer's Disease Sequencing Project, the UK Biobank, the Finnish health registry, and the US Million Veterans Program. Risk of AD was evaluated through case-control logistic regression analyses. Data were analyzed between January 2023 and March 2024. Genetic data available from high-density single-nucleotide variant microarrays and whole-genome sequencing and summary statistics for multitissue expression and protein quantitative trait loci available from published studies were included, enabling follow-up genetic colocalization analyses. A total of 1 629 863 eligible participants were selected from referred and volunteer samples, 477 596 of whom were excluded for analysis exclusion criteria. The number of participants who declined to participate in original studies was not available.

Main Outcome and Measures

Risk of AD, reported as odds ratios (ORs) with 95% CIs. Associations were considered at X chromosome-wide (P < 1 × 10-5) and genome-wide (P < 5 × 10-8) significance. Primary analyses are nonstratified, while secondary analyses evaluate sex-stratified effects.

Results

Analyses included 1 152 284 participants of non-Hispanic White, European ancestry (664 403 [57.7%] female and 487 881 [42.3%] male), including 138 558 individuals with AD. Six independent genetic loci passed X chromosome-wide significance, with 4 showing support for links between the genetic signal for AD and expression of nearby genes in brain and nonbrain tissues. One of these 4 loci passed conservative genome-wide significance, with its lead variant centered on an intron of SLC9A7 (OR, 1.03; 95% CI, 1.02-1.04) and colocalization analyses prioritizing both the SLC9A7 and nearby CHST7 genes. Of these 6 loci, 4 displayed evidence for escape from X chromosome inactivation with regard to AD risk.

Conclusion and Relevance

This large-scale XWAS of AD identified the novel SLC9A7 locus. SLC9A7 regulates pH homeostasis in Golgi secretory compartments and is anticipated to have downstream effects on amyloid β accumulation. Overall, this study advances our knowledge of AD genetics and may provide novel biological drug targets. The results further provide initial insights into elucidating the role of the X chromosome in sex-based differences in AD.

Summary statistics knockoffs inference with family-wise error rate control

Testing multiple hypotheses of conditional independence with provable error rate control is a fundamental problem with various applications. To infer conditional independence with family-wise error rate (FWER) control when only summary statistics of marginal dependence are accessible, we adopt GhostKnockoff to directly generate knockoff copies of summary statistics and propose a new filter to select features conditionally dependent on the response. In addition, we develop a computationally efficient algorithm to greatly reduce the computational cost of knockoff copies generation without sacrificing power and FWER control. Experiments on simulated data and a real dataset of Alzheimer's disease genetics demonstrate the advantage of the proposed method over existing alternatives in both statistical power and computational efficiency.

Rare genetic variation in fibronectin 1 (FN1) protects against APOEε4 in Alzheimer's disease

The risk of developing Alzheimer's disease (AD) significantly increases in individuals carrying the APOEε4 allele. Elderly cognitively healthy individuals with APOEε4 also exist, suggesting the presence of cellular mechanisms that counteract the pathological effects of APOEε4; however, these mechanisms are unknown. We hypothesized that APOEε4 carriers without dementia might carry genetic variations that could protect them from developing APOEε4-mediated AD pathology. To test this, we leveraged whole-genome sequencing (WGS) data in the National Institute on Aging Alzheimer's Disease Family Based Study (NIA-AD FBS), Washington Heights/Inwood Columbia Aging Project (WHICAP), and Estudio Familiar de Influencia Genetica en Alzheimer (EFIGA) cohorts and identified potentially protective variants segregating exclusively among unaffected APOEε4 carriers. In homozygous unaffected carriers above 70 years old, we identified 510 rare coding variants. Pathway analysis of the genes harboring these variants showed significant enrichment in extracellular matrix (ECM)-related processes, suggesting protective effects of functional modifications in ECM proteins. We prioritized two genes that were highly represented in the ECM-related gene ontology terms, (FN1) and collagen type VI alpha 2 chain (COL6A2) and are known to be expressed at the blood-brain barrier (BBB), for postmortem validation and in vivo functional studies. An independent analysis in a large cohort of 7185 APOEε4 homozygous carriers found that rs140926439 variant in FN1 was protective of AD (OR = 0.29; 95% CI [0.11, 0.78], P = 0.014) and delayed age at onset of disease by 3.37 years (95% CI [0.42, 6.32], P = 0.025). The FN1 and COL6A2 protein levels were increased at the BBB in APOEε4 carriers with AD. Brain expression of cognitively unaffected homozygous APOEε4 carriers had significantly lower FN1 deposition and less reactive gliosis compared to homozygous APOEε4 carriers with AD, suggesting that FN1 might be a downstream driver of APOEε4-mediated AD-related pathology and cognitive decline. To validate our findings, we used zebrafish models with loss-of-function (LOF) mutations in fn1b-the ortholog for human FN1. We found that fibronectin LOF reduced gliosis, enhanced gliovascular remodeling, and potentiated the microglial response, suggesting that pathological accumulation of FN1 could impair toxic protein clearance, which is ameliorated with FN1 LOF. Our study suggests that vascular deposition of FN1 is related to the pathogenicity of APOEε4, and LOF variants in FN1 may reduce APOEε4-related AD risk, providing novel clues to potential therapeutic interventions targeting the ECM to mitigate AD risk.

APOE loss-of-function variants: Compatible with longevity and associated with resistance to Alzheimer's disease pathology

The ε4 allele of apolipoprotein E (APOE) is the strongest genetic risk factor for sporadic Alzheimer's disease (AD). Knockdown of ε4 may provide a therapeutic strategy for AD, but the effect of APOE loss of function (LoF) on AD pathogenesis is unknown. We searched for APOE LoF variants in a large cohort of controls and patients with AD and identified seven heterozygote carriers of APOE LoF variants. Five carriers were controls (aged 71-90 years), one carrier was affected by progressive supranuclear palsy, and one carrier was affected by AD with an unremarkable age at onset of 75 years. Two APOE ε3/ε4 controls carried a stop-gain affecting ε4: one was cognitively normal at 90 years and had no neuritic plaques at autopsy; the other was cognitively healthy at 79 years, and lumbar puncture at 76 years showed normal levels of amyloid. These results suggest that ε4 drives AD risk through the gain of abnormal function and support ε4 knockdown as a viable therapeutic option.

Machine Learning Approach to Identify Case-Control Studies on ApoE Gene Mutations Linked to Alzheimer’s Disease in Italy

Background: An application of artificial intelligence is machine learning, which allows computer programs to learn and create data. Methods: In this work, we aimed to evaluate the performance of the MySLR machine learning platform, which implements the Latent Dirichlet Allocation (LDA) algorithm in the identification and screening of papers present in the literature that focus on mutations of the apolipoprotein E (ApoE) gene in Italian Alzheimer’s Disease patients. Results: MySLR excludes duplicates and creates topics. MySLR was applied to analyze a set of 164 scientific publications. After duplicate removal, the results allowed us to identify 92 papers divided into two relevant topics characterizing the investigated research area. Topic 1 contains 70 papers, and topic 2 contains the remaining 22. Despite the current limitations, the available evidence suggests that articles containing studies on Italian Alzheimer’s Disease (AD) patients were 65.22% (n = 60). Furthermore, the presence of papers about mutations, including single nucleotide polymorphisms (SNPs) ApoE gene, the primary genetic risk factor of AD, for the Italian population was 5.4% (n = 5). Conclusion: The results show that the machine learning platform helped to identify case-control studies on ApoE gene mutations, including SNPs, but not only conducted in Italy.

Controlled Variable Selection from Summary Statistics Only? A Solution via GhostKnockoffs and Penalized Regression

Identifying which variables do influence a response while controlling false positives pervades statistics and data science. In this paper, we consider a scenario in which we only have access to summary statistics, such as the values of marginal empirical correlations between each dependent variable of potential interest and the response. This situation may arise due to privacy concerns, e.g., to avoid the release of sensitive genetic information. We extend GhostKnockoffs He et al. [2022] and introduce variable selection methods based on penalized regression achieving false discovery rate (FDR) control. We report empirical results in extensive simulation studies, demonstrating enhanced performance over previous work. We also apply our methods to genome-wide association studies of Alzheimer's disease, and evidence a significant improvement in power.

APOE Genotype and Alzheimer Disease Risk Across Age, Sex, and Population Ancestry

Importance

Apolipoprotein E (APOE)*2 and APOE*4 are, respectively, the strongest protective and risk-increasing, common genetic variants for late-onset Alzheimer disease (AD), making APOE status highly relevant toward clinical trial design and AD research broadly. The associations of APOE genotypes with AD are modulated by age, sex, race and ethnicity, and ancestry, but these associations remain unclear, particularly among racial and ethnic groups understudied in the AD and genetics research fields.

Objective

To assess the stratified associations of APOE genotypes with AD risk across sex, age, race and ethnicity, and global population ancestry.

Design, Setting, Participants

This genetic association study included case-control, family-based, population-based, and longitudinal AD-related cohorts that recruited referred and volunteer participants. Data were analyzed between March 2022 and April 2023. Genetic data were available from high-density, single-nucleotide variant microarrays, exome microarrays, and whole-exome and whole-genome sequencing. Summary statistics were ascertained from published AD genetic studies.

Main Outcomes and Measures

The main outcomes were risk for AD (odds ratios [ORs]) and risk of conversion to AD (hazard ratios [HRs]), with 95% CIs. Risk for AD was evaluated through case-control logistic regression analyses. Risk of conversion to AD was evaluated through Cox proportional hazards regression survival analyses.

Results

Among 68 756 unique individuals, analyses included 21 852 East Asian (demographic data not available), 5738 Hispanic (68.2% female; mean [SD] age, 75.4 [8.8] years), 7145 non-Hispanic Black (hereafter referred to as Black) (70.8% female; mean [SD] age, 78.4 [8.2] years), and 34 021 non-Hispanic White (hereafter referred to as White) (59.3% female; mean [SD] age, 77.0 [9.1] years) individuals. There was a general, stepwise pattern of ORs for APOE*4 genotypes and AD risk across race and ethnicity groups. Odds ratios for APOE*34 and AD risk attenuated following East Asian (OR, 4.54; 95% CI, 3.99-5.17),White (OR, 3.46; 95% CI, 3.27-3.65), Black (OR, 2.18; 95% CI, 1.90-2.49) and Hispanic (OR, 1.90; 95% CI, 1.65-2.18) individuals. Similarly, ORs for APOE*22+23 and AD risk attenuated following White (OR, 0.53, 95% CI, 0.48-0.58), Black (OR, 0.69, 95% CI, 0.57-0.84), and Hispanic (OR, 0.89; 95% CI, 0.72-1.10) individuals, with no association for Hispanic individuals. Deviating from the global pattern of ORs, APOE*22+23 was not associated with AD risk in East Asian individuals (OR, 0.97; 95% CI, 0.77-1.23). Global population ancestry could not explain why Hispanic individuals showed APOE associations with less pronounced AD risk compared with Black and White individuals. Within Black individuals, decreased global African ancestry or increased global European ancestry showed a pattern of APOE*4 dosage associated with increasing AD risk, but no such pattern was apparent for APOE*2 dosage with AD risk. The sex-by-age-specific interaction effect of APOE*34 among White individuals (higher risk in women) was reproduced but shifted to ages 60 to 70 years (OR, 1.48; 95% CI, 1.10-2.01) and was additionally replicated in a meta-analysis of Black individuals and Hispanic individuals (OR, 1.72; 95% CI, 1.01-2.94).

Conclusion and Relevance

Through recent advances in AD-related genetic cohorts, this study provided the largest-to-date overview of the association of APOE with AD risk across age, sex, race and ethnicity, and population ancestry. These novel insights are critical to guide AD clinical trial design and research.

APOE loss-of-function variants: Compatible with longevity and associated with resistance to Alzheimer's Disease pathology

The ε4 allele of apolipoprotein E (APOE) is the strongest genetic risk factor for sporadic Alzheimer's Disease (AD). Knockdown of this allele may provide a therapeutic strategy for AD, but the effect of APOE loss-of-function (LoF) on AD pathogenesis is unknown. We searched for APOE LoF variants in a large cohort of older controls and patients with AD and identified six heterozygote carriers of APOE LoF variants. Five carriers were controls (ages 71-90) and one was an AD case with an unremarkable age-at-onset between 75-79. Two APOE ε3/ε4 controls (Subjects 1 and 2) carried a stop-gain affecting the ε4 allele. Subject 1 was cognitively normal at 90+ and had no neuritic plaques at autopsy. Subject 2 was cognitively healthy within the age range 75-79 and underwent lumbar puncture at between ages 75-79 with normal levels of amyloid. The results provide the strongest human genetics evidence yet available suggesting that ε4 drives AD risk through a gain of abnormal function and support knockdown of APOE ε4 or its protein product as a viable therapeutic option.

GhostKnockoff inference empowers identification of putative causal variants in genome-wide association studies

Recent advances in genome sequencing and imputation technologies provide an exciting opportunity to comprehensively study the contribution of genetic variants to complex phenotypes. However, our ability to translate genetic discoveries into mechanistic insights remains limited at this point. In this paper, we propose an efficient knockoff-based method, GhostKnockoff, for genome-wide association studies (GWAS) that leads to improved power and ability to prioritize putative causal variants relative to conventional GWAS approaches. The method requires only Z-scores from conventional GWAS and hence can be easily applied to enhance existing and future studies. The method can also be applied to meta-analysis of multiple GWAS allowing for arbitrary sample overlap. We demonstrate its performance using empirical simulations and two applications: (1) a meta-analysis for Alzheimer's disease comprising nine overlapping large-scale GWAS, whole-exome and whole-genome sequencing studies and (2) analysis of 1403 binary phenotypes from the UK Biobank data in 408,961 samples of European ancestry. Our results demonstrate that GhostKnockoff can identify putatively functional variants with weaker statistical effects that are missed by conventional association tests.