Alzheimer's Disease Genetics Consortium Research Articles

An X Chromosome Transcriptome Wide Association Study Implicates ARMCX6 in Alzheimer's Disease.

The X chromosome is often omitted in disease association studies despite containing thousands of genes that may provide insight into well-known sex differences in the risk of Alzheimer's disease (AD). To model the expression of X chromosome genes and evaluate their impact on AD risk in a sex-stratified manner. Using elastic net, we evaluated multiple modeling strategies in a set of 175 whole blood samples and 126 brain cortex samples, with whole genome sequencing and RNA-seq data. SNPs (MAF > 0.05) within the cis-regulatory window were used to train tissue-specific models of each gene. We apply the best models in both tissues to sex-stratified summary statistics from a meta-analysis of Alzheimer's Disease Genetics Consortium (ADGC) studies to identify AD-related genes on the X chromosome. Across different model parameters, sample sex, and tissue types, we modeled the expression of 217 genes (95 genes in blood and 135 genes in brain cortex). The average model R2 was 0.12 (range from 0.03 to 0.34). We also compared sex-stratified and sex-combined models on the X chromosome. We further investigated genes that escaped X chromosome inactivation (XCI) to determine if their genetic regulation patterns were distinct. We found ten genes associated with AD at p < 0.05, with only ARMCX6 in female brain cortex (p=0.008) nearing the significance threshold after adjusting for multiple testing (α=0.002). We optimized the expression prediction of X chromosome genes, applied these models to sex-stratified AD GWAS summary statistics, and identified one putative AD risk gene, ARMCX6.

Genomic assessment of early‐onset AD identifies novel risk loci and an incomplete genetic overlap with late‐onset AD

AbstractBackgroundEarly Onset Alzheimer Disease (EOAD, age at onset [AAO] &lt; = 65) is a severe form of AD, often occurring when patients are still caring for children or adults. Despite this, most genetic studies of EOAD have focused on autosomal dominant forms, and there is little understanding of similarities and differences between early and late onset forms of AD (LOAD, [AAO]&gt;65). To address this, we present a genome‐wide association study of non‐Mendelian EOAD and compare it to LOAD GWAS.MethodsPrimary single‐variant analyses were performed using two additive logistic regression for case‐control models (full: SNPs, PCs, sex and APOE‐ɛ4 dosage as covariates; reduced: SNPs and PCs); secondary models included PCs+sex and PCs+APOE. Models were applied on data derived from the Alzheimer Disease Genetics Consortium (ADGC), composed by unrelated individuals. Samples varied from 1293‐1459 cases and 8894‐9366 controls for EOAD, and 8795‐9508 cases and 9702‐10273 controls for LOAD, depending on the considered model. The SNP heritability (h2) and genetic correlation (rg) were estimated using LD score regression. Gene‐based and pathway analysis (gene sets from Msigdb‐v7.0) were performed using FUMA/MAGMA‐v1.6.ResultsWe identified two novel loci associated with EOAD: Chromosome 4 (full model: chr4:102027610, P = 2.98×10−08, near PPP3CA gene) and Chromosome 12 (PCs+sex model: rs117001070, P = 2.11×10−08, intergenic between LINC02444 and LINC02882). We additionally confirmed BIN1 and APOE in the EOAD subset. Several known genes also showed a significant association with LOAD. Heritability analyses showed higher h2 values for EOAD (h2 = 0.29, 0.29, 0.29 and 0.27) than LOAD (h2 = 0.19, 0.16, 0.19, and 0.15), for full, reduced, PCs+sex and PCs+APOE models respectively. Genetic correlation showed moderate but incomplete genetic correlation between EOAD and LOAD. Other than APOE (p = 3.89×10−12), gene‐based tests showed nominal association for EOAD C8orf44‐SGK3 (p = 9.95×10−06), SGK3 (p = 2.01×10−05) and HIST1H2AC (p = 3.04×10−05), and plausible pathways such as HDL remodeling (p = 5.50×10−05) and positive regulation of cholesterol efflux (p = 9.79×10−05). These pathways were nominally associated with LOAD (p = 0.04 and p = 4.05×10−03).ConclusionsWe identified two novel loci associated to EOAD not previously reported by LOAD studies. Heritability and genetic correlation results suggest that the genetic etiology of EOAD has an incomplete genetic overlap with LOAD.

Variants near X‐Chromosome Genes NLGN4X and PTCHD1 are Significantly Associated with Alzheimer’s Disease Risk

AbstractBackgroundGenome‐wide association studies (GWAS) for AD have identified numerous associated loci but have been focused only on the autosomes. We sought to identify novel AD‐associated genes on X‐Chromosome (X‐Chr).MethodWe evaluated the association of AD with X‐Chr variants in GWAS datasets assembled by the Alzheimer Disease Genetics Consortium including 26,322 non‐Hispanic White (NHW) individuals using logistic regression models with covariates for age, sex and principal components of ancestry. The number of independent SNPs in X‐Chr was calculated by linkage disequilibrium pruning yielding a study‐wide significance threshold of P&lt;7.58×10−7. Genes near SNPs surpassing this threshold were further evaluated by expression quantitative trait locus (eQTL) in the BRAINEAC database and differential gene expression analyses using RNA‐seq data obtained from autopsied brains contained (GEO: GSE44772 and GSE33000). We also tested association of AD with missense or splicing variants with minor allele count (MAC)≥10 in the same genes using GENESIS and whole exome sequencing (WES) data obtained by the Alzheimer Disease Sequencing Project (ADSP) for 11,172 NHWs.ResultWe identified novel associations with variants near NLGN4X (rs34056759; minor allele frequency [MAF] = 0.15, OR = 1.15, P = 5.14×10−7) and PTCHD1 (rs5970663; MAF = 0.30, OR = 1.13, P = 8.26×10−7). The rs5970663 risk allele C was significantly associated with increased expression of PTCHD1 in the medulla region, and PTCHD1 expression was significantly higher in AD cases than in controls in the prefrontal cortex (P&lt;2.1×10−8) and visual cortex (P&lt;2.3×10−4). Rs34056759 was not associated with NLGN4X expression, but NLGN4X expression was significantly lower in AD cases compared to controls in the prefrontal cortex (P&lt;8.7×10−7). A rare PTCHD1 variant (rs201933353; MAF = 0.001) was nominally associated with AD among 6,318 women (OR = 3.72, P = 0.06).ConclusionWe identified novel associations of AD with X‐Chr variants near NLGN4X and PTCHD1, loci that were previously associated with autism spectrum disorder. We plan to replicate, fine‐map, and annotate the candidate loci identified in this study using large whole genome sequencing ADSP datasets.

A haptoglobin (HP) structural variant alters the effect of APOE alleles on Alzheimer's disease.

Haptoglobin (HP) is an antioxidant of apolipoprotein E (APOE), and previous reports have shown HP binds with APOE and amyloid beta (Aβ) to aid its clearance. A common structural variant of the HP gene distinguishes it into two alleles: HP1 and HP2. HP genotypes were imputed in 29 cohorts from the Alzheimer's Disease Genetics Consortium (N=20,512). Associations between the HP polymorphism and Alzheimer's disease (AD) risk and age of onset through APOE interactions were investigated using regression models. The HP polymorphism significantly impacts AD risk in European-descent individuals (and in meta-analysis with African-descent individuals) by modifying both the protective effect of APOE ε2 and the detrimental effect of APOE ε4. The effect is particularly significant among APOE ε4 carriers. The effect modification of APOE by HP suggests adjustment and/or stratification by HP genotype is warranted when APOE risk is considered. Our findings also provided directions for further investigations on potential mechanisms behind this association.

Identification of novel genes for age-at-onset of Alzheimer's disease by combining quantitative and survival trait analyses.

Our understanding of the genetic predisposition for age-at-onset (AAO) of Alzheimer's disease (AD) is limited. Here, we sought to identify genes modifying AAO and examined whether any have sex-specific effects. Genome-wide association analysis were performed on imputed genetic data of 9219 AD cases and 10,345 controls from 20 cohorts of the Alzheimer's Disease Genetics Consortium. AAO was modeled from cases directly and as a survival outcome. We identified 11 genome-wide significant loci (P<5×10-8 ), including six known AD-risk genes and five novel loci, UMAD1, LUZP2, ARFGEF2, DSCAM, and 4q25, affecting AAO of AD. Additionally, 39 suggestive loci showed strong association. Twelve loci showed sex-specific effects on AAO including CD300LG and MLX/TUBG2 for females and MIR4445 for males. Genes that influence AAO of AD are excellent therapeutic targets for delaying onset of AD. Several loci identified include genes with promising functional implications for AD.

Explainable machine learning aggregates polygenic risk scores and electronic health records for Alzheimer’s disease prediction

Alzheimer’s disease (AD) is the most common late-onset neurodegenerative disorder. Identifying individuals at increased risk of developing AD is important for early intervention. Using data from the Alzheimer Disease Genetics Consortium, we constructed polygenic risk scores (PRSs) for AD and age-at-onset (AAO) of AD for the UK Biobank participants. We then built machine learning (ML) models for predicting development of AD, and explored feature importance among PRSs, conventional risk factors, and ICD-10 codes from electronic health records, a total of > 11,000 features using the UK Biobank dataset. We used eXtreme Gradient Boosting (XGBoost) and SHapley Additive exPlanations (SHAP), which provided superior ML performance as well as aided ML model explanation. For participants age 40 and older, the area under the curve for AD was 0.88. For subjects of age 65 and older (late-onset AD), PRSs were the most important predictors. This is the first observation that PRSs constructed from the AD risk and AAO play more important roles than age in predicting AD. The ML model also identified important predictors from EHR, including urinary tract infection, syncope and collapse, chest pain, disorientation and hypercholesterolemia, for developing AD. Our ML model improved the accuracy of AD risk prediction by efficiently exploring numerous predictors and identified novel feature patterns.

African ancestry GWAS of dementia in a large military cohort identifies significant risk loci.

While genome wide association studies (GWASs) of Alzheimer's Disease (AD) in European (EUR) ancestry cohorts have identified approximately 83 potentially independent AD risk loci, progress in non-European populations has lagged. In this study, data from the Million Veteran Program (MVP), a biobank which includes genetic data from more than 650,000 US Veteran participants, was used to examine dementia genetics in an African descent (AFR) cohort. A GWAS of Alzheimer's disease and related dementias (ADRD), an expanded AD phenotype including dementias such as vascular and non-specific dementia that included 4012 cases and 18,435 controls age 60+ in AFR MVP participants was performed. A proxy dementia GWAS based on survey-reported parental AD or dementia (n = 4385 maternal cases, 2256 paternal cases, and 45,970 controls) was also performed. These two GWASs were meta-analyzed, and then subsequently compared and meta-analyzed with the results from a previous AFR AD GWAS from the Alzheimer's Disease Genetics Consortium (ADGC). A meta-analysis of common variants across the MVP ADRD and proxy GWASs yielded GWAS significant associations in the region of APOE (p = 2.48 × 10-101), in ROBO1 (rs11919682, p = 1.63 × 10-8), and RNA RP11-340A13.2 (rs148433063, p = 8.56 × 10-9). The MVP/ADGC meta-analysis yielded additional significant SNPs near known AD risk genes TREM2 (rs73427293, p = 2.95 × 10-9), CD2AP (rs7738720, p = 1.14 × 10-9), and ABCA7 (rs73505251, p = 3.26 × 10-10), although the peak variants observed in these genes differed from those previously reported in EUR and AFR cohorts. Of the genes in or near suggestive or genome-wide significant associated variants, nine (CDA, SH2D5, DCBLD1, EML6, GOPC, ABCA7, ROS1, TMCO4, and TREM2) were differentially expressed in the brains of AD cases and controls. This represents the largest AFR GWAS of AD and dementia, finding non-APOE GWAS-significant common SNPs associated with dementia. Increasing representation of AFR participants is an important priority in genetic studies and may lead to increased insight into AD pathophysiology and reduce health disparities.

A Million Veteran Program GWAS of Alzheimer’s Disease and Related Dementias in African Americans Identifies Multiple Genome‐Wide Significant Dementia Risk Loci

AbstractBackgroundWhile genome wide association studies (GWASs) of Alzheimer’s Disease (AD) in European ancestry cohorts (EUR) have identified &gt;70 potentially independent AD risk loci, progress in non‐European populations has lagged. In this study, data from US Veterans of African ancestry (AFR) from the Million Veteran Program (MVP) biorepository were used to identify genetic factors associated with AD and related dementias (ADRD).MethodUsing the VA electronic medical record, 4,012 AFR ADRD cases (algorithm includes codes for AD and dementias such as vascular dementia) and 18,435 dementia‐free AFR controls were identified for inclusion in the GWAS. A GWAS was also performed using proxy dementia based on survey‐reported AFR participant parental history of AD or other dementia that included 4,385 maternal cases, 2,256 paternal cases, and 45,970 controls. Meta‐analysis of the case control and proxy results for was performed and then subsequently combined with results from a AFR AD GWAS conducted by the Alzheimer Disease Genetics Consortium (ADGC; Kunkle et al., 2021).ResultThe AFR MVP GWAS (case‐control + proxy) yielded genome‐wide significant associations with several loci including APOE (p=2.48x10‐101), ROBO1 (rs11919682, p=1.63x10‐8), and RP11‐340A13.2 (rs148433063, p=8.56x10‐9. The ROBO1 and RP11‐240A13.2 SNPs were not replicated in the ADGC sample (missing or not significant). The meta‐analysis of MVP and ADGC GWAS results (Table 1) additionally identified genome significant SNPs near known AD risk genes TREM2 (rs73427293, p=2.95x10‐9), CD2AP (rs7738720, p=1.14x10‐9), and ABCA7 (rs73505251, p=3.26x10‐10). Comparison of expression of genes near variants with suggestive evidence of association (p&lt;5x10‐7) in the brains of EUR AD cases and controls from Framingham Heart Study, Religious Orders Study/Rush Memory and Aging Project, Mayo‐Mount Sinai Brain Bank, and Boston University Brain Bank (Panitch et al.) identified several differentially expressed genes including EML6.ConclusionThe AFR dementia GWAS presented here is the largest to date. It increased the number of genome wide significant common‐variant dementia loci reported in AFR cohorts from two to six, including genes previously implicated in EUR cohorts. Increasing representation of AFRs is an important priority in genetic studies and may lead to increased insight into AD pathophysiology and reduce health disparities.

A large‐scale genome‐wide association study of early‐onset Alzheimer disease

AbstractBackgroundEarly‐onset (≤65 years old) Alzheimer Disease (EOAD), which accounts for approximately 10% of all AD cases, is sometimes equated with dominantly inherited forms of AD (i.e., caused by known mutations in APP, PSEN1 or PSEN2). However, 90% of EOAD cases remain unexplained by these variants. This project aims to identify novel EOAD‐associated variants through a large‐scale, multi‐ethnic genome‐wide association (GWA) study.MethodWe have leveraged the Alzheimer Disease Genetics Consortium (ADGC) GWAS dataset that is aligned to GRCh38 and imputed with the TOPMed imputation server. This includes 39 studies from non‐Hispanic Whites (NHW; CA = 3,202; CO = 5,782), 17 from African American (CA = 139; CO = 253), six from Asian (CA = 133; CO = 474), and seven from Hispanic (CA = 199; CO = 59) cohorts. Only the younger (≤65yo) and older participants (&gt;80yo) were included for analysis as cases and controls, respectively. The genetic background was confirmed using PCA. We are performing single‐variant analysis to identify variants that confer a higher risk for EOAD.ResultOur preliminary single‐variant association analysis (using sex, PC1‐PC10 as covariates) emerging from the largest NHW cohort (CA = 3,202; CO = 5,782) have revealed nine different loci that were associated with EOAD at a suggestive P‐value&lt;1×10−6. The strongest signal from our analysis was in the APOE region. Additionally, we identified three novel genome‐wide significant (P‐value &lt;5×10−8) loci of interest at HLA‐DRB9 (rs9268852, P‐value = 2.9×10−9, OR = 1.22); FAM86B3P (rs73199790, P‐value = 3×10−11, OR = 1.32); and MSMB (rs2075895, P‐value = 2.83×10−16, OR = 1.26); while also confirming the previously identified loci at MS4A4A (rs7108663, P‐value = 1.13×10−10, OR = 0.80) and CR1 (rs6656401, P‐value = 2.54×10−12, OR = 1.32). We will perform qualitative associations on the other three ethnicities to confirm the transferability of these signals.ConclusionOur single‐variant analysis has identified three novel loci associated with EOAD. This study is the largest genetic screening for risk and protective variants in EOAD and will be instrumental to identifying novel variants and pathways implicated in unexplained EOAD. Our future analysis includes gene‐based association and pathway analyses per ethnicity to identify variants that confer a higher risk for EOAD, followed by transethnic analysis to validate the comprehensive nature of these risk factors. Additionally, we have also generated GWAS data from the Knight‐ADRC participants, which will be merged with the ADGC datasets for our future analysis.

Genome-wide association and multi-omics studies identify MGMT as a novel risk gene for Alzheimer's disease among women.

Variants in the tau gene (MAPT) region are associated with breast cancer in women and Alzheimer's disease (AD) among persons lacking apolipoprotein E ε4 (ε4-). To identify novel genes associated with tau-related pathology, we conducted two genome-wide association studies (GWAS) for AD, one among 10,340 ε4- women in the Alzheimer's Disease Genetics Consortium (ADGC) and another in 31 members (22 women) of a consanguineous Hutterite kindred. We identified novel associations of AD with MGMT variants in the ADGC (rs12775171, odds ratio [OR] = 1.4, P=4.9×10-8 ) and Hutterite (rs12256016 and rs2803456, OR = 2.0, P=1.9×10-14 ) datasets. Multi-omics analyses showed that the most significant and largest number of associations among the single nucleotide polymorphisms (SNPs), DNA-methylated CpGs, MGMT expression, and AD-related neuropathological traits were observed among women. Furthermore, promoter capture Hi-C analyses revealed long-range interactions of the MGMT promoter with MGMT SNPs and CpG sites. These findings suggest that epigenetically regulated MGMT expression is involved in AD pathogenesis, especially in women.

Genome-wide association and multi-omics studies identify MGMT as a novel risk gene for Alzheimer disease among women.

Increasing evidence suggests epidemiological and pathological links between Alzheimer disease (AD) and breast cancer (BC). The 17q21 locus, including MAPT encoding tau, has been identified in genome-wide association studies (GWAS) for BC in women and AD among persons lacking the APOE ε4 allele (ε4-). We hypothesized that other genes involved in tau-related pathology can be identified by GWAS in women stratified by ε4 status. We conducted a GWAS for AD among ε4- women (3,400 AD and 6,940 controls) in Alzheimer's Disease Genetics Consortium (ADGC) datasets and separately analyzed whole genome sequencing data obtained from five AD cases (all women; ε4+=4) and 26 controls (17 women, 9 men; ε4+=9) in a Hutterite kindred from South Dakota. Association tests were performed using logistic regression with covariates for age and principle components of ancestry in the ADGC datasets and for age, sex, and ε4 in the Hutterites. Potential mechanisms of top-ranked findings were investigated by integrating omics data including gene expression, DNA-methylation (DNAm), and AD-related neuropathological traits in subsets stratified by sex and ε4 status in the Religious Orders Study and Memory and Aging Project (ROSMAP; 339 AD and 229 controls). We identified genome-wide significant associations with one novel locus MGMT in each sample (ADGC: rs12775171, odds ratio [OR]=1.4, P=4.9x10-8 ; Hutterites: rs12256016, OR=2.0, P=1.9x10-14 ). Rs12775171 showed no evidence of association with AD risk in ε4+ women or in men in the ADGC datasets (P>0.05). Rs12775171 is not in linkage disequilibrium with rs12256016. At least one of these SNPs was significantly associated with increased DNAm levels near/within MGMT (best association pair: rs11016864 [proxy for rs12256016] with cg01419164; P=4.0x10-4 in women), which were also significantly associated with lower MGMT expression (P=0.03) and worse tau burden in all AD cohorts (neurofibrillary tangle; P=6.0x10-3 ), women (Braak stage; P=4.0x10-3 ), and ε4- women (neurofibrillary tangle; P=0.03), but not among ε4+ women, all men, or ε4+/ε4- men in ROSMAP. We identified a novel association of AD with the interaction of sex and MGMT. Our omics analysis confirmed that epigenetically regulated expression of MGMT (involved in DNA methylation damage repair) is significantly associated with tau in women.

A large-scale, whole genome sequencing study of unexplained early-onset Alzheimer disease.

Genomic studies of Alzheimer disease (AD) have primarily focused on non-Hispanic White (NHW) participants affected by the late-onset form (LOAD; onset age: >65), or the study of early onset AD (EOAD; onset age <=65) cases showing Mendelian inheritance patterns associated with mutations in the APP, PSEN1 and PSEN2 genes. However, mutations in these three genes explain ∼10% of EOAD cases. There are no large-scale efforts to collect and study EOAD cases not explained by these genes, despite this unexplained category accounting for ∼90% of EOAD cases. To address this, we aim to identify additional EOAD-associated variants, genes and pathways through a large-scale whole-genome sequencing (WGS) study of unexplained EOAD. We will include cases from several AD cohorts, including the Resource for Early-onset Alzheimer Disease Research (READR), the Knight-ADRC at Washington University, the Alzheimer's Disease Genetics Consortium (ADGC), and others. Generating and harmonizing a dataset of 200 non-Hispanic White (NHW) and Caribbean Hispanic (CH) multiplex EOAD families, over 5,400 EOAD singletons and over 13,000 unrelated, cognitive controls, all with WGS, this project will yield the largest EOAD genomics dataset to-date, improving statistical power for variant identification and allowing us to assess the impact of specific factors such as APOE genotype, vascular risk factors, and neuropsychiatric comorbidities. The inclusion of a large set of Hispanic families and singletons allows the examination of EOAD risk in a significantly understudied population. Analyses will comprise both linkage and association-based approaches, analyses of polygenic and ancestry effects, and a thorough examination of neurocognitive, neuropsychiatric and cardiovascular endophenotypes. When completed this study will point to novel genetic contributors to EOAD, shed light on the mechanisms of AD and facilitate the development of novel prediction models and therapeutics. Sampling, phenotyping and sequencing analysis protocols will be complementary to and compatible with the existing LOAD genomics resources, such as the Alzheimer Disease Sequencing Project (ADSP) and related studies. This phenotypic and genomic consistency, together with the use of existing AD infrastructure (NIAGADS), allows for immediate integration with the leading efforts on LOAD, enabling rapid large-scale investigation of a variety of additional critical AD genomics hypotheses.

ADRC GWAS supplement: An NCRAD initiative.

In collaboration with the Alzheimer's Disease Genetics Consortium (ADGC), the National Centralized Repository for Alzheimer's Disease and Related Dementias (NCRAD) sends DNA from Alzheimer's Disease Research Centers (ADRC) subjects for Genome-Wide Association Studies (GWAS); however, only a subset of ADRC subjects have been assayed. NCRAD has initiated an effort to supplement this initiative, as well as generate GWAS for targeted additional studies. NCRAD sends DNA to the Children's Hospital of Philadelphia (CHOP) Center for Applied Genomics (CAG). CHOP performs GWAS using the Illumina Infinium Global Screening Array BeadChip v2.0 (Illumina, Inc., San Diego, CA). Genotype data is processed using the same pipeline as the ADGC GWAS. Data for ADRC subjects will be returned to the contributing ADRC, and all GWAS data will be posted at The National Institute on Aging Genetics of Alzheimer's Disease Data Storage Site (NIAGADS), available for download by approved researchers. NCRAD is also creating online catalogues with demographic and clinical data for access by approved researchers, which can be used in conjunction with the GWAS data for research. To date, NCRAD has sent 4,910 non-ADGC ADRC samples with the National Alzheimer's Coordinating Center Uniform Data Set to CHOP for GWAS; genotyping is complete and data has been transferred to NIAGADS for processing. Additionally, NCRAD shipped 1,655 samples from the Amyloid Neuroimaging and Genetics Initiative (ANGI), and 275 samples for the 90+ Study; genotyping is complete for both studies and data has been transferred to NIAGADS. NCRAD will also send samples from the Longitudinal Early-Onset Alzheimer's Disease Study (LEADS) for GWAS this spring, with the future goal that all LEADS subjects will have GWAS data available. NCRAD is completing online catalogues with demographic and clinical data for ANGI and the 90+ study. NCRAD has made a significant contribution to expand the focus and scope of AD genetics research by supplementing the ADRC GWAS cohort, as well as providing GWAS for several unique and high impact studies. We anticipate that our continued efforts will bolster current resources, supporting the goal to identify new loci associated with early and late onset AD.

NIA genetics of Alzheimer's disease data storage site (NIAGADS): 2021 update.

NIAGADS is a national genomics data repository that facilitates access of genotypic and sequencing data to qualified investigators for the study of the genetics of Alzheimer's disease (AD) and related neurological diseases. Collaborations with large consortia and centers such as the Alzheimer's Disease Genetics Consortium (ADGC), Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE) Consortium, the Alzheimer's Disease Sequencing Project (ADSP), and the Genome Center for Alzheimer's Disease (GCAD) allow NIAGADS to lead the effort in managing large AD datasets that can be easily accessed and fully utilized by the research community. NIAGADS is supported by National Institute on Aging (NIA) under a cooperative agreement. All data derived from NIA funded AD genetics studies are expected to be deposited in NIAGADS or another NIA approved site. NIAGADS manages a Data Sharing Service (DSS) that facilitates the deposition and sharing of genomic data and association results with approved users in the neurodegenerative research community. In addition, researchers are able to freely use the NIAGADS Alzheimer's Genomics Database (www.niagads.org/genomics/) to search annotation resources that link published AD studies to AD-relevant sequence features and genome-wide annotations. As of January 2021, NIAGADS houses 74 datasets comprised of >90,000 samples including GWAS, sequencing, gene expression, annotations, deep phenotypes, and summary statistics. Qualified investigators can retrieve ADSP sequencing data with ease and flexibility through the NIAGADS DSS. As of February 2021, the ADSP and other contributing studies have completed whole exome sequencing (WES) of 20,504 samples and whole-genome sequencing (WGS) of 16,908 samples. Raw WES and WGS files, quality controlled VCF files, and phenotype data files are available via qualified access. The next round of sequencing currently underway will generate around 18,000 additional genomes to be released at the end of 2021. NIAGADS is a rich resource for AD researchers, with the goal of facilitating advances in Alzheimer's genetics research. By housing datasets from many projects and institutions, NIAGADS enables AD researchers to meet their research goals more efficiently. Datasets, guidelines, and new features are available on our website at https://www.niagads.org.

Genetic association with clusters of persons with TDP-43 proteinopathy.

TAR-DNA binding protein 43 kDa (TDP-43) proteinopathy is a neurodegenerative pathology present in up to 50% of aged persons with dementia. Recently, a classification system for limbic-predominant age-related TDP-43 encephalopathy neuropathologic changes (LATE-NC) was proposed. In previous work, we reported that brains with TDP-43 proteinopathy comprise four distinct clusters, based on clinical features and comorbid neuropathologies (PMID 32797255). All included participants (N=495) had autopsy-confirmed TDP-43 proteinopathy. Cluster 1 (n=103) contained mostly cases with a pathologic diagnosis of frontotemporal lobar degeneration (FTLD-TDP). Cluster 2 (n=71) was distinguished by relatively "pure" LATE-NC (i.e., no/low ADNC). Both Clusters 3 (n=114) and 4 (n=207) had high levels of ADNC + LATE-NC; however, Cluster 4 featured earlier disease onset and swift disease progression without FTLD clinical features. Here, we investigate differences in genetic features between ADNC and LATE-NC subclusters. Participant data were drawn from the previously analyzed sample (N=495), which comprised of autopsied research volunteers who were followed at Alzheimer's Disease Research Centers and appear in the National Alzheimer's Coordinating Center database. Alzheimer's Disease Genetics Consortium (ADGC) data were linked to the initial sample, and we examined the association of single-nucleotide polymorphisms (SNPs) identified as a late-onset AD associated SNP by the International Genomics of Alzheimer's Disease Project genome wide-association study in LATE-NC clusters. ADGC data were available for 124/495 participants. Of these, 89 were in Cluster 3 and 4 (ADNC+LATE-NC) and 20 in Cluster 2 (pure LATE-NC). The SNPs rs190982 located in MEF2C on Chromosome 5 and rs11218343 located in SORL1 on Chromosome 11 were significantly associated with cluster membership. The G allele of rs190982 and the C allele of rs11218343 were risk SNPs for Cluster 2. Two SNPs may be differentially associated with ADNC+LATE-NC vs. pure LATE-NC. Given that TDP-43 proteinopathy is common in aged brains, identifying genetic factors is a crucial step toward developing disease-modifying therapies. Although our final sample size was small, these results might help point to unsuspected pathways that influence ADNC and LATE-NC phenotypes.

The ε4‐bearing TOMM40‐APOE‐APOC1 haplotype but not the ε4 allele confers an exceptionally high risk of Alzheimer’s disease

The APOE gene harboring the ε2 and ε4 alleles encoded by minor alleles of rs7412 and rs429358 single nucleotide polymorphisms (SNPs), respectively, is one of the most studied genes in humans. Despite that, the problem of whether the associations between these alleles and the late-onset Alzheimer's disease (referred to as AD) are due to these alleles themselves remains controversial. This problem is complicated by an inherent heterogeneity in genetic predisposition to AD, which is common for complex diseases in late life.We used a comprehensive, two-stage strategy to gain insights into a more homogeneous genetic architecture of AD in subjects of European ancestry in the TOMM40-APOE-APOC1 region, emphasizing the role of the ε4 allele. First, we examined differences in linkage disequilibrium (LD) patterns in AD-affected and unaffected subjects who carry or do not carry the ε4 or ε2 allele to prioritize SNP pairs according to the LD differences. Second, rs429358 and SNPs from the top pair were selected to construct compound genotypes and associate them with AD. The discovery analyses used a mega-sample of 2,673 AD-affected and 16,246 cognitively normal subjects from four independent studies. The landmark Alzheimer's Disease Genetics Consortium initiative cohort (3,662 AD-cases and 1,541 controls) was used for validation.The top LD difference between the most unfavorable AD-affected ε2-free sample and the most favorable AD- and ε4-free sample was for rs2075650 (TOMM40) and rs12721046 (APOC1) SNPs, Δr=77.6% (p<10-100 ). The meta-analysis results showed the exceptionally high risk of AD for carriers of the full minor allele homozygote comprised of rs429358, rs2075650, and rs12721046 SNPs, odds ratio=18.1, p=1.96×10-77 . This risk was 4.37-fold (p=1.34×10-3 ) higher than that for carriers of two copies of the ε4 alleles who do not have minor alleles of rs2075650 and rs12721046.This study provides compelling evidence that the strongest genetic predisposition to AD in the APOE region is credited to a haplotype comprised of the ε4 allele and TOMM40 (rs2075650) and APOC1 (rs12721046) minor alleles, but not to the ε4 allele itself. This more homogeneous polygenic profile is promising for assessing subjects at an exceptionally high risk of AD.

Heritability analyses show partial genetic overlap between (non‐Mendelian) early and late onset Alzheimer disease due to an intriguing APOE effect

Alzheimer disease (AD) is a degenerative brain disease, being the most common cause of progressive dementia and listed as the sixth leading cause of mortality in the USA. It is often described as either early onset (EOAD, age at onset, [AAO] <= 65) or late onset (LOAD, [AAO]>65). Non-Mendelian EOAD (not monogenic; nmEOAD) has irregular inheritance patterns and fluctuating AAO, characteristics also present in LOAD cases. There is still a lack of evidence in the literature depicting the similarities (if any) between nmEOAD and LOAD forms, being unclear how much genetic etiology is shared by the two forms of AD. To shed light to this question, a genome-wide association study (GWAS) and heritability analyses of nmEOAD and LOAD were performed.Genetic data on 21,622 individuals from the Alzheimer Disease Genetics Consortium (ADGC) were used: (1,476 nmEOAD, 9,695 LOAD and 10,451 control). Single-variant association analyses were performed using logistic regression under two models: (1) ancestry plus SNP, and (2) ancestry, sex, APOE dosage, and SNP. nmEOAD and LOAD were considered separately. LD score regression was used to estimate the SNP heritability (h2 ) and genetic correlation (rg), considering two additional models: (3) ancestry, sex, and SNP and (4) ancestry, APOE dosage and SNP.Several known candidate genes confirmed for LOAD along with novel regions associated with immune and cell-signaling pathways in nmEOAD models. Gene based tests showed significant association for APOE gene (Chr19): nmEOAD (p=3.89x10-16 and p=4.29x10-12 ) and LOAD (p=1.07x10-65 and p=1.12x10-14 ), models (1) and (2) respectively. Heritability analyses showed higher h2 values for EOAD (h2 =0.24, 0.23, 0.25 and 0.24) than LOAD (h2 =0.18, 0.14, 0.18, and 0.14) for models (1) to (4) respectively. Genetic correlation showed moderate genetic overlap between EOAD and LOAD only for models: (2) rg=0.35 (p=0.0283) and (4) rg=0.34 (p=0.0261).GWAS and heritability analysis suggest that the genetic etiology of EOAD has a noncomplete genetic overlap with LOAD, with a moderate overlap when APOE dosage is modeled and a minimal overlap otherwise (APOE effect). The results also suggest a stronger polygenic effect in EOAD than LOAD, confirming the need for additional genomics efforts in nmEOAD.

Alzheimer’s disease risk prediction using automated machine learning

AbstractBackgroundAlzheimer’s disease (AD) is the most common late‐onset neurodegenerative disease. About 5.4 million Americans are living with AD. Unfortunately, there is no cure for AD at present, which makes early prediction crucial. Identifying individuals at increased risk of AD provides a better chance of benefiting from treatments. Risk prediction models are typically based on a limited number of predictors possibly with sub‐optimal performance. Here, we explore a state‐of‐the‐art automated machine learning (AutoML) framework for AD risk prediction, which can handle hundreds of predictors, including non‐traditional variables, with automatic feature engineering and model selection.MethodWe developed an AutoML model that aggregates polygenic risk scores (PRSs) and baseline individual characteristics (e.g., non‐genetic factors) for predicting AD. The PRSs were derived using summary statistics of the genome‐wide association studies from the Alzheimer Disease Genetics Consortium (ADGC) dataset (n = 19,918). The model was applied to 455,233 participants in UKBB without AD at baseline to predict development of AD at the final observation (n=1,452 developed AD). Our model was based on the H2O AutoML, an intelligent algorithm that can automatically select hyperparameters, tune ensembles of ML models, and carry out model assessment.ResultThe area under the receiver operating characteristic curve (AUC) for AD risk prediction was over 0.86. Polygenic risk scores ranked only second to age in feature importance. Furthermore, our AutoML model identified predictors that are not typically considered in traditional prediction models, such as an individual’s overall health rating and usual walking pace.ConclusionOur AutoML model improves the accuracy of AD risk prediction by efficiently exploring numerous predictors and ensemble models while greatly reducing manual coding hours. Furthermore, AutoML uncovered novel predictors for AD.

Definitive roles of TOMM40-APOE-APOC1 variants in the Alzheimer's risk

Despite advances, the roles of genetic variants from the APOE-harboring 19q13.32 region in Alzheimer's disease (AD) remain controversial. We leverage a comprehensive approach to gain insights into a more homogeneous genetic architecture of AD in this region. We use a sample of 2,673 AD-affected and 16,246 unaffected subjects from 4 studies and validate our main findings in the landmark Alzheimer's Disease Genetics Consortium cohort (3,662 AD-cases and 1,541 controls). We report the remarkably high excesses of the AD risk for carriers of the ε4 allele who also carry minor alleles of rs2075650 (TOMM40) and rs12721046 (APOC1) polymorphisms compared to carriers of their major alleles. The exceptionally high 4.37-fold (p=1.34 × 10−3) excess was particularly identified for the minor allele homozygotes. The beneficial and adverse variants were significantly depleted and enriched, respectively, in the AD-affected families. This study provides compelling evidence for the definitive roles of the APOE-TOMM40-APOC1 variants in the AD risk.

Mechanism for the protective effect of APOE ε2 against Alzheimer disease is linked to tau and the classical complement pathway

AbstractBackgroundRecent reports suggest that the rare APOE Christchurch mutation and APOE2 allele protect against Alzheimer’s disease (AD) pathology by reducing the burden of tau pathology. However, the mechanism(s) underlying the APOE2 protective effect linking to tau is largely unknown.MethodWe conducted a genome‐wide association study (GWAS) for AD among 2,120 APOE2 carriers from the Alzheimer Disease Genetics Consortium (ADGC), and prioritized by gene network analysis and methylation profiling of CpG sites in the locus using the Religious Orders Study and Memory and Aging Project (ROSMAP) data. We also performed differential gene expression analysis among APOE2/3 subjects in prefrontal cortex tissue from 44 AD and 75 control brains of the Framingham Heart Study (FHS), Boston University Alzheimer’s Disease Center (BUADC), and ROSMAP. We validated the relationships between protein levels of the top‐ranked genes and levels of tau and amyloid beta proteins in prefrontal cortex area of 193 FHS/BUADC brains.ResultOf the significant genetic associations with AD (P &lt; 10−6), PPP2CB was the key node in APOE2‐related gene networks and contained the most significant CpG site (P = 5.2 × 10−4) located 2,814 bp upstream of the top‐ranked GWAS variant. The complement genes C4A and C4B were the top‐ranked differentially expressed genes (P &lt; 2 × 10−4) with reduced expression in APOE2/3 controls compared in APOE2/3 cases or other carriers. Furthermore, the level of PPP2CB protein was significantly associated with increased levels of C4B protein (P = 8.2 × 10−7), and the C4A level was associated with increased pTau231/tTau ratio (P = 0.006). Protein levels of C4A, C4B, and PPP2CB were not associated with amyloid beta 42 (Aβ42) or pTau181/tTau ratio. We were unable to confirm previous reports on the role of C1q, since the C1q level was not associated with any of these phenotypes.ConclusionThis is the first to report the molecular link between a tau phosphatase and the classical complement pathway, especially C4, as a mechanism for the ε2 protective effect against AD. Our findings may lead to the development of novel therapeutic targets for AD by mimicking the neuroprotective effect of APOE2.