Bridging the gap between statistical and biological epistasis in Alzheimer's disease

Genome-Wide Epistasis Study of Cerebrospinal Fluid Hyperphosphorylated Tau in ADNI Cohort

Alzheimer's disease (AD) is the main cause of dementia worldwide, and the genetic mechanism of which is not yet fully understood. Much evidence has accumulated over the past decade to suggest that after the first large-scale genome-wide association studies (GWAS) were conducted, the problem of "missing heritability" in AD is still a great challenge. Epistasis has been considered as one of the main causes of "missing heritability" in AD, which has been largely ignored in human genetics. The focus of current genome-wide epistasis studies is usually on single nucleotide polymorphisms (SNPs) that have significant individual effects, and the amount of heritability explained by which was very low. Moreover, AD is characterized by progressive cognitive decline and neuronal damage, and some studies have suggested that hyperphosphorylated tau (P-tau) mediates neuronal death by inducing necroptosis and inflammation in AD. Therefore, this study focused on identifying epistasis between two-marker interactions at marginal main effects across the whole genome using cerebrospinal fluid (CSF) P-tau as quantitative trait (QT). We sought to detect interactions between SNPs in a multi-GPU based linear regression method by using age, gender, and clinical diagnostic status (cds) as covariates. We then used the STRING online tool to perform the PPI network and identify two-marker epistasis at the level of gene-gene interaction. A total of 758 SNP pairs were found to be statistically significant. Particularly, between the marginal main effect SNP pairs, highly significant SNP-SNP interactions were identified, which explained a relatively high variance at the P-tau level. In addition, 331 AD-related genes were identified, 10 gene-gene interaction pairs were replicated in the PPI network. The identified gene-gene interactions and genes showed associations with AD in terms of neuroinflammation and neurodegeneration, neuronal cells activation and brain development, thereby leading to cognitive decline in AD, which is indirectly associated with the P-tau pathological feature of AD and in turn supports the results of this study. Thus, the results of our study might be beneficial for explaining part of the "missing heritability" of AD.

A Novel Multitasking Ant Colony Optimization Method for Detecting Multiorder SNP Interactions

MOTIVATION

Linear or nonlinear interactions of multiple single-nucleotide polymorphisms (SNPs) play an important role in understanding the genetic basis of complex human diseases. However, combinatorial analytics in high-dimensional space makes it extremely challenging to detect multiorder SNP interactions. Most classic approaches can only perform one task (for detecting k-order SNP interactions) in each run. Since prior knowledge of a complex disease is usually not available, it is difficult to determine the value of k for detecting k-order SNP interactions.

METHODS

A novel multitasking ant colony optimization algorithm (named MTACO-DMSI) is proposed to detect multiorder SNP interactions, and it is divided into two stages: searching and testing. In the searching stage, multiple multiorder SNP interaction detection tasks (from 2nd-order to kth-order) are executed in parallel, and two subpopulations that separately adopt the Bayesian network-based K2-score and Jensen-Shannon divergence (JS-score) as evaluation criteria are generated for each task to improve the global search capability and the discrimination ability for various disease models. In the testing stage, the G test statistical test is adopted to further verify the authenticity of candidate solutions to reduce the error rate.

RESULT

Three multiorder simulated disease models with different interaction effects and three real age-related macular degeneration (AMD), rheumatoid arthritis (RA) and type 1 diabetes (T1D) datasets were used to investigate the performance of the proposed MTACO-DMSI. The experimental results show that the MTACO-DMSI has a faster search speed and higher discriminatory power for diverse simulation disease models than traditional single-task algorithms. The results on real AMD data and RA and T1D datasets indicate that MTACO-DMSI has the ability to detect multiorder SNP interactions at a genome-wide scale. Availability and implementation: https://github.com/shouhengtuo/MTACO-DMSI/.

Maximal Information Coefficient-Based Testing to Identify Epistasis in Case-Control Association Studies

Interactions between genetic variants (epistasis) are ubiquitous in the model system and can significantly affect evolutionary adaptation, genetic mapping, and precision medical efforts. In this paper, we proposed a method for epistasis detection, called EpiMIC (epistasis detection through a maximal information coefficient (MIC)). MIC is a promising bivariate dependence measure explicitly designed for rapidly exploring various function types equally and for interpreting and comparing them on the same scale. Most epistasis detection approaches make assumptions about the form of the association between genetic variants, resulting in limited statistical performance. Based on the notion that if two SNPs do not interact, their joint distribution in all samples and in only cases should not be substantially different. We developed a statistic that utilizes the difference of MIC as a signal of epistasis and combined it with a permutation resampling strategy to estimate the empirical distribution of our statistic. Results of simulation and real-world data set showed that EpiMIC outperformed previous approaches for identifying epistasis at varying degrees of heredity.

Genetic Bases of Complex Traits: From Quantitative Trait Loci to Prediction

Conceived as a general introduction to the book, this chapter is a reminder of the core concepts of genetic mapping and molecular marker-based prediction. It provides an overview of the principles and the evolution of methods for mapping the variation of complex traits, and methods for QTL-based prediction of human disease risk and animal and plant breeding value. The principles of linkage-based and linkage disequilibrium-based QTL mapping methods are described in the context of the simplest, single-marker, methods. Methodological evolutions are analysed in relation with their ability to account for the complexity of the genotype-phenotype relations. Main characteristics of the genetic architecture of complex traits, drawn from QTL mapping works using large populations of unrelated individuals, are presented. Methods combining marker-QTL association data into polygenic risk score that captures part of an individual's susceptibility to complex diseases are reviewed. Principles of best linear mixed model-based prediction of breeding value in animal- and plant-breeding programs using phenotypic and pedigree data, are summarized and methods for moving from BLUP to marker-QTL BLUP are presented. Factors influencing the additional genetic progress achieved by using molecular data and rules for their optimization are discussed.

Interplay between stress-related genes may influence Alzheimer's disease development: The results of genetic interaction analyses of human data

Emerging evidence from experimental and clinical research suggests that stress-related genes may play key roles in AD development. The fact that genome-wide association studies were not able to detect a contribution of such genes to AD indicates the possibility that these genes may influence AD non-linearly, through interactions of their products. In this paper, we selected two stress-related genes (GCN2/EIF2AK4 and APP) based on recent findings from experimental studies which suggest that the interplay between these genes might influence AD in humans. To test this hypothesis, we evaluated the effects of interactions between SNPs in these two genes on AD occurrence, using the Health and Retirement Study data on white indidividuals. We found several interacting SNP-pairs whose associations with AD remained statistically significant after correction for multiple testing. These findings emphasize the importance of nonlinear mechanisms of polygenic AD regulation that cannot be detected in traditional association studies. To estimate collective effects of multiple interacting SNP-pairs on AD, we constructed a new composite index, called Interaction Polygenic Risk Score, and showed that its association with AD is highly statistically significant. These results open a new avenue in the analyses of mechanisms of complex multigenic AD regulation.

Emerging roles of rare and low-frequency genetic variants in type 1 diabetes mellitus

Type 1 diabetes mellitus (T1DM) is defined as an autoimmune disorder and has enormous complexity and heterogeneity. Although its precise pathogenic mechanisms are obscure, this disease is widely acknowledged to be precipitated by environmental factors in individuals with genetic susceptibility. To date, the known susceptibility loci, which have mostly been identified by genome-wide association studies, can explain 80%–85% of the heritability of T1DM. Researchers believe that at least a part of its missing genetic component is caused by undetected rare and low-frequency variants. Most common variants have only small to modest effect sizes, which increases the difficulty of dissecting their functions and restricts their potential clinical application. Intriguingly, many studies have indicated that rare and low-frequency variants have larger effect sizes and play more significant roles in susceptibility to common diseases, including T1DM, than common variants do. Therefore, better recognition of rare and low-frequency variants is beneficial for revealing the genetic architecture of T1DM and for providing new and potent therapeutic targets for this disease. Here, we will discuss existing challenges as well as the great significance of this field and review current knowledge of the contributions of rare and low-frequency variants to T1DM.

Roles of interacting stress-related genes in lifespan regulation: insights for translating experimental findings to humans

AIM

Experimental studies provided numerous evidence that caloric/dietary restriction may improve health and increase the lifespan of laboratory animals, and that the interplay among molecules that sense cellular stress signals and those regulating cell survival can play a crucial role in cell response to nutritional stressors. However, it is unclear whether the interplay among corresponding genes also plays a role in human health and lifespan.

METHODS

Literature about roles of cellular stressors have been reviewed, such as amino acid deprivation, and the integrated stress response (ISR) pathway in health and aging. Single nucleotide polymorphisms (SNPs) in two candidate genes (GCN2/EIF2AK4 and CHOP/DDIT3) that are closely involved in the cellular stress response to amino acid starvation, have been selected using information from experimental studies. Associations of these SNPs and their interactions with human survival in the Health and Retirement Study data have been estimated. The impact of collective associations of multiple interacting SNP pairs on survival has been evaluated, using a recently developed composite index: the SNP-specific Interaction Polygenic Risk Score (SIPRS).

RESULTS

Significant interactions have been found between SNPs from GCN2/EIF2AK4 and CHOP/DDI3T genes that were associated with survival 85+ compared to survival between ages 75 and 85 in the total sample (males and females combined) and in females only. This may reflect sex differences in genetic regulation of the human lifespan. Highly statistically significant associations of SIPRS [constructed for the rs16970024 (GCN2/EIF2AK4) and rs697221 (CHOP/DDIT3)] with survival in both sexes also been found in this study.

CONCLUSION

Identifying associations of the genetic interactions with human survival is an important step in translating the knowledge from experimental to human aging research. Significant associations of multiple SNPxSNP interactions in ISR genes with survival to the oldest old age that have been found in this study, can help uncover mechanisms of multifactorial regulation of human lifespan and its heterogeneity.

Alzheimer Disease Pathology-Associated Polymorphism in a Complex Variable Number of Tandem Repeat Region Within the MUC6 Gene, Near the AP2A2 Gene

We found evidence of late-onset Alzheimer disease (LOAD)-associated genetic polymorphism within an exon of Mucin 6 (MUC6) and immediately downstream from another gene: Adaptor Related Protein Complex 2 Subunit Alpha 2 (AP2A2). PCR analyses on genomic DNA samples confirmed that the size of the MUC6 variable number tandem repeat (VNTR) region was highly polymorphic. In a cohort of autopsied subjects with quantitative digital pathology data (n = 119), the size of the polymorphic region was associated with the severity of pTau pathology in neocortex. In a separate replication cohort of autopsied subjects (n = 173), more pTau pathology was again observed in subjects with longer VNTR regions (p = 0.031). Unlike MUC6, AP2A2 is highly expressed in human brain. AP2A2 expression was lower in a subset analysis of brain samples from persons with longer versus shorter VNTR regions (p = 0.014 normalizing with AP2B1 expression). Double-label immunofluorescence studies showed that AP2A2 protein often colocalized with neurofibrillary tangles in LOAD but was not colocalized with pTau proteinopathy in progressive supranuclear palsy, or with TDP-43 proteinopathy. In summary, polymorphism in a repeat-rich region near AP2A2 was associated with neocortical pTau proteinopathy (because of the unique repeats, prior genome-wide association studies were probably unable to detect this association), and AP2A2 was often colocalized with neurofibrillary tangles in LOAD.

The MUC6/AP2A2 Locus and Its Relevance to Alzheimer's Disease: A Review

We recently reported evidence of Alzheimer's disease (AD)-linked genetic variation within the mucin 6 (MUC6) gene on chromosome 11p, nearby the adaptor-related protein complex 2 subunit alpha 2 (AP2A2) gene. This locus has interesting features related to human genomics and clinical research. MUC6 gene variants have been reported to potentially influence viral-including herpesvirus-immunity and the gut microbiome. Within the MUC6 gene is a unique variable number of tandem repeat (VNTR) region. We discovered an association between MUC6 VNTR repeat expansion and AD pathologic severity, particularly tau proteinopathy. Here, we review the relevant literature. The AD-linked VNTR polymorphism may also influence AP2A2 gene expression. AP2A2 encodes a polypeptide component of the adaptor protein complex, AP-2, which is involved in clathrin-coated vesicle function and was previously implicated in AD pathogenesis. To provide background information, we describe some key knowledge gaps in AD genetics research. The "missing/hidden heritability problem" of AD is highlighted. Extensive portions of the human genome, including the MUC6 VNTR, have not been thoroughly evaluated due to limitations of existing high-throughput sequencing technology. We present and discuss additional data, along with cautionary considerations, relevant to the hypothesis that MUC6 repeat expansion influences AD pathogenesis.

The combined risk effect among BIN1, CLU, and APOE genes in Alzheimer's disease

Genome-wide associations studies (GWAS) are detecting new variants associated with late-onset of Alzheimer’s disease (LOAD), a multifactorial neurodegenerative disorder. The variants rs744373 BIN1, rs11136000 CLU and rs3764650 ABCA7 uncovered by GWAS led to different AD pathways, such as metabolism, trafficking and endocytosis of lipids and inflammation. However, most of the association studies did not replicate these variants with significance. This could be due to a small power effect evident when these variants are tested independently with LOAD. Therefore, we aimed to investigate whether the combination of different variants would additively modify the risk of association with LOAD that is observed in GWAS. We performed an association study testing pairwise variants in metabolism, trafficking and endocytosis of lipid (rs429358 and rs7412 APOE, rs744373 BIN1, rs3764650 ABCA7 and rs11136000 CLU) pathways with LOAD in samples from southeastern Brazil. Our data suggest a risk effect for LOAD between APOE with CLU and APOE with BIN1 genes.

Failure to detect synergy between variants in transferrin and hemochromatosis and Alzheimer's disease in large cohort

Alzheimer's disease (AD) is the most common cause of dementia and, despite decades of effort, there is no effective treatment. In the last decade, many association studies have identified genetic markers that are associated with AD status. Two of these studies suggest that an epistatic interaction between variants rs1049296 in the transferrin (TF) gene and rs1800562 in the homeostatic iron regulator (HFE) gene, commonly known as hemochromatosis, is in genetic association with AD. TF and HFE are involved in the transport and regulation of iron in the brain, and disrupting these processes exacerbates AD pathology through increased neurodegeneration and oxidative stress. However, by using a significantly larger data set from the Alzheimer's Disease Genetics Consortium, we fail to detect an association between TF rs1049296 or HFE rs1800562 with AD risk (TF rs1049296 p = 0.38 and HFE rs1800562 p = 0.40). In addition, logistic regression with an interaction term and a synergy factor analysis both failed to detect epistasis between TF rs1049296 and HFE rs1800562 (SF = 0.94; p = 0.48) in AD cases. Each of these analyses had sufficient statistical power (power > 0.99), suggesting that previously reported associations may be the result of more complex epistatic interactions, genetic heterogeneity, or false-positive associations because of limited sample sizes.

Systematic analysis of dark and camouflaged genes reveals disease-relevant genes hiding in plain sight

BACKGROUND

The human genome contains "dark" gene regions that cannot be adequately assembled or aligned using standard short-read sequencing technologies, preventing researchers from identifying mutations within these gene regions that may be relevant to human disease. Here, we identify regions with few mappable reads that we call dark by depth, and others that have ambiguous alignment, called camouflaged. We assess how well long-read or linked-read technologies resolve these regions.

RESULTS

Based on standard whole-genome Illumina sequencing data, we identify 36,794 dark regions in 6054 gene bodies from pathways important to human health, development, and reproduction. Of these gene bodies, 8.7% are completely dark and 35.2% are ≥ 5% dark. We identify dark regions that are present in protein-coding exons across 748 genes. Linked-read or long-read sequencing technologies from 10x Genomics, PacBio, and Oxford Nanopore Technologies reduce dark protein-coding regions to approximately 50.5%, 35.6%, and 9.6%, respectively. We present an algorithm to resolve most camouflaged regions and apply it to the Alzheimer's Disease Sequencing Project. We rescue a rare ten-nucleotide frameshift deletion in CR1, a top Alzheimer's disease gene, found in disease cases but not in controls.

CONCLUSIONS

While we could not formally assess the association of the CR1 frameshift mutation with Alzheimer's disease due to insufficient sample-size, we believe it merits investigating in a larger cohort. There remain thousands of potentially important genomic regions overlooked by short-read sequencing that are largely resolved by long-read technologies.

How to increase our belief in discovered statistical interactions via large-scale association studies?

The understanding that differences in biological epistasis may impact disease risk, diagnosis, or disease management stands in wide contrast to the unavailability of widely accepted large-scale epistasis analysis protocols. Several choices in the analysis workflow will impact false-positive and false-negative rates. One of these choices relates to the exploitation of particular modelling or testing strategies. The strengths and limitations of these need to be well understood, as well as the contexts in which these hold. This will contribute to determining the potentially complementary value of epistasis detection workflows and is expected to increase replication success with biological relevance. In this contribution, we take a recently introduced regression-based epistasis detection tool as a leading example to review the key elements that need to be considered to fully appreciate the value of analytical epistasis detection performance assessments. We point out unresolved hurdles and give our perspectives towards overcoming these.

Male-specific epistasis between WWC1 and TLN2 genes is associated with Alzheimer's disease

Systematic epistasis analyses in multifactorial disorders are an important step to better characterize complex genetic risk structures. We conducted a hypothesis-free sex-stratified genome-wide screening for epistasis contributing to Alzheimer's disease (AD) susceptibility. We identified a statistical epistasis signal between the single nucleotide polymorphisms rs3733980 and rs7175766 that was associated with AD in males (genome-wide significant p_{Bonferroni-corrected}=0.0165). This signal pointed toward the genes WW and C2 domain containing 1, aka KIBRA; 5q34 and TLN2 (talin 2; 15q22.2). Gene-based meta-analysis in 3 independent consortium data sets confirmed the identified interaction: the most significant (p_{meta-Bonferroni-corrected}=9.02*10^-3) was for the single nucleotide polymorphism pair rs1477307 and rs4077746. In functional studies, WW and C2 domain containing 1, aka KIBRA and TLN2 coexpressed in the temporal cortex brain tissue of AD subjects (β=0.17, 95% CI 0.04 to 0.30, p=0.01); modulated Tau toxicity in Drosophila eye experiments; colocalized in brain tissue cells, N2a neuroblastoma, and HeLa cell lines; and coimmunoprecipitated both in brain tissue and HEK293 cells. Our finding points toward new AD-related pathways and provides clues toward novel medical targets for the cure of AD.

Improving the Gene Ontology Resource to Facilitate More Informative Analysis and Interpretation of Alzheimer's Disease Data

The analysis and interpretation of high-throughput datasets relies on access to high-quality bioinformatics resources, as well as processing pipelines and analysis tools. Gene Ontology (GO, geneontology.org) is a major resource for gene enrichment analysis. The aim of this project, funded by the Alzheimer's Research United Kingdom (ARUK) foundation and led by the University College London (UCL) biocuration team, was to enhance the GO resource by developing new neurological GO terms, and use GO terms to annotate gene products associated with dementia. Specifically, proteins and protein complexes relevant to processes involving amyloid-beta and tau have been annotated and the resulting annotations are denoted in GO databases as 'ARUK-UCL'. Biological knowledge presented in the scientific literature was captured through the association of GO terms with dementia-relevant protein records; GO itself was revised, and new GO terms were added. This literature biocuration increased the number of Alzheimer's-relevant gene products that were being associated with neurological GO terms, such as 'amyloid-beta clearance' or 'learning or memory', as well as neuronal structures and their compartments. Of the total 2055 annotations that we contributed for the prioritised gene products, 526 have associated proteins and complexes with neurological GO terms. To ensure that these descriptive annotations could be provided for Alzheimer's-relevant gene products, over 70 new GO terms were created. Here, we describe how the improvements in ontology development and biocuration resulting from this initiative can benefit the scientific community and enhance the interpretation of dementia data.

Navβ2 knockdown improves cognition in APP/PS1 mice by partially inhibiting seizures and APP amyloid processing

Voltage-gated sodium channels beta 2 (Navβ2, encoded by SCN2B) is a substrate of β-site amyloid precursor protein cleaving enzyme 1 (BACE1) and regulates cell surface expression of channels in neurons. Previous studies reported enhanced Navβ2 processing by BACE1 in Alzheimer’s disease (AD) model and patients. We investigated whether changes in Navβ2 expression affect neuronal seizure and amyloid precursor protein (APP) processing in an AD mouse model. Our study used eight-month-old APP/presenilin 1 (PS1) mice and transgenic Navβ2 knockdown [by 61% vs. wild type (WT)] APP/PS1 mice (APP/PS1/Navβ2-kd), with age-matched WT and Navβ2 knockdown (Navβ2-kd) mice as controls. We found that Navβ2 knockdown in APP/PS1 mice partially reversed the abnormal Navβ2 cleavage and the changes in intracellular and total Nav1.1α expression. It also restored sodium currents density in hippocampal neurons and neuronal activity, as indicated by EEG tracing; improved Morris water maze performance; and shifted APP amyloidogenic metabolism towards non-amyloidogenic processing. There were no differences in these indicators between WT and Navβ2-kd mice. These results suggest Navβ2 knockdown may be a promising strategy for treating AD.

Genetic epistasis regulates amyloid deposition in resilient aging

INTRODUCTION

The brain-derived neurotrophic factor (BDNF) interacts with important genetic Alzheimer's disease (AD) risk factors. Specifically, variants within the SORL1 gene determine BDNF's ability to reduce amyloid β (Aβ) in vitro. We sought to test whether functional BDNF variation interacts with SORL1 genotypes to influence expression and downstream AD-related processes in humans.

METHODS

We analyzed postmortem brain RNA sequencing and neuropathological data for 441 subjects from the Religious Orders Study/Memory and Aging Project and molecular and structural neuroimaging data for 1285 subjects from the Alzheimer's Disease Neuroimaging Initiative.

RESULTS

We found one SORL1 RNA transcript strongly regulated by SORL1-BDNF interactions in elderly without pathological AD and showing stronger associations with diffuse than neuritic Aβ plaques. The same SORL1-BDNF interactions also significantly influenced Aβ load as measured with [¹⁸F]Florbetapir positron emission tomography.

DISCUSSION

Our results bridge the gap between risk and resilience factors for AD, demonstrating interdependent roles of established SORL1 and BDNF functional genotypes.

A novel approach for multi-SNP GWAS and its application in Alzheimer's disease

BACKGROUND

Genome-wide association studies (GWAS) have effectively identified genetic factors for many diseases. Many diseases, including Alzheimer's disease (AD), have epistatic causes, requiring more sophisticated analyses to identify groups of variants which together affect phenotype.

RESULTS

Based on the GWAS statistical model, we developed a multi-SNP GWAS analysis to identify pairs of variants whose common occurrence signaled the Alzheimer's disease phenotype.

CONCLUSIONS

Despite not having sufficient data to demonstrate significance, our preliminary experimentation identified a high correlation between GRIA3 and HLA-DRB5 (an AD gene). GRIA3 has not been previously reported in association with AD, but is known to play a role in learning and memory.

Presenilin E318G variant and Alzheimer's disease risk: the Cache County study

Background

Alzheimer's disease is the leading cause of dementia in the elderly and the third most common cause of death in the United States. A vast number of genes regulate Alzheimer’s disease, including Presenilin 1 (PSEN1). Multiple studies have attempted to locate novel variants in the PSEN1 gene that affect Alzheimer's disease status. A recent study suggested that one of these variants, PSEN1 E318G (rs17125721), significantly affects Alzheimer's disease status in a large case–control dataset, particularly in connection with the APOEε4 allele.

Methods

Our study looks at the same variant in the Cache County Study on Memory and Aging, a large population-based dataset. We tested for association between E318G genotype and Alzheimer’s disease status by running a series of Fisher’s exact tests. We also performed logistic regression to test for an additive effect of E318G genotype on Alzheimer’s disease status and for the existence of an interaction between E318G and APOEε4.

Results

In our Fisher’s exact test, it appeared that APOEε4 carriers with an E318G allele have slightly higher risk for AD than those without the allele (3.3 vs. 3.8); however, the 95 % confidence intervals of those estimates overlapped completely, indicating non-significance. Our logistic regression model found a positive but non-significant main effect for E318G (p = 0.895). The interaction term between E318G and APOEε4 was also non-significant (p = 0.689).

Conclusions

Our findings do not provide significant support for E318G as a risk factor for AD in APOEε4 carriers. Our calculations indicated that the overall sample used in the logistic regression models was adequately powered to detect the sort of effect sizes observed previously. However, the power analyses of our Fisher’s exact tests indicate that our partitioned data was underpowered, particularly in regards to the low number of E318G carriers, both AD cases and controls, in the Cache county dataset. Thus, the differences in types of datasets used may help to explain the difference in effect magnitudes seen. Analyses in additional case–control datasets will be required to understand fully the effect of E318G on Alzheimer's disease status.

Interaction between variants in CLU and MS4A4E modulates Alzheimer's disease risk

INTRODUCTION

Ebbert et al. reported gene-gene interactions between rs11136000-rs670139 (CLU-MS4A4E) and rs3865444-rs670139 (CD33-MS4A4E). We evaluate these interactions in the largest data set for an epistasis study.

METHODS

We tested interactions using 3837 cases and 4145 controls from Alzheimer's Disease Genetics Consortium using meta-analyses and permutation analyses. We repeated meta-analyses stratified by apolipoprotein E (APOE) ε4 status, estimated combined odds ratio (OR) and population attributable fraction (cPAF), and explored causal variants.

RESULTS

Results support the CLU-MS4A4E interaction and a dominant effect. An association between CLU-MS4A4E and APOE ε4 negative status exists. The estimated synergy factor, OR, and cPAF for rs11136000-rs670139 are 2.23, 2.45, and 8.0, respectively. We identified potential causal variants.

DISCUSSION

We replicated the CLU-MS4A4E interaction in a large case-control series and observed APOE ε4 and possible dominant effect. The CLU-MS4A4E OR is higher than any Alzheimer's disease locus except APOE ε4, APP, and TREM2. We estimated an 8% decrease in Alzheimer's disease incidence without CLU-MS4A4E risk alleles and identified potential causal variants.