NG00119 – Health and Retirement Study Genotype Data 2006-2012

We will use the HRS data to pursue two complementary strategies. One is the discovery of particular genetic polymorphisms associated with social-science outcomes. Because the effect of an individual genetic polymorphism on the outcome is likely to be very small, the HRS sample is too small, taken alone, to be used to discover new associations. Hence, we will pursue this strategy with HRS data in conjunction with other datasets that we have organized in the Social Science Genetic Association Consortium (SSGAC; www.thessgac.org). Our second strategy focuses on exploiting the uniquely rich social-science data in the HRS. We will conduct analyses that will shed light on the genetic architecture of a range of social-science outcomes. We will apply statistical methods that use the information contained in the dense SNP data taken as a whole and are thus well-powered in a sample size such as that of the HRS. Our specific aims are: 1. To incorporate data from the HRS into ongoing meta-GWAS efforts from the SSGAC for a range of social-science outcomes, such as educational attainment, and personality. 2. To continue to include HRS in the future releases of the Polygenic Index (PGI) Repository. PGIs (aka polygenic scores) are summaries of a person's genetic predisposition to a particular trait. HRS was included in the first release of the Repository, for which we created PGIs for 47 phenotypes in 11 datasets, which were returned to the datasets to be shared with users according to the datasets’ own data sharing procedures. We will regularly update the existing PGIs and add new phenotypes as larger GWAS and better methodologies become available. Details on the Repository can be found in Becker et al. (2021, Resource profile and user guide of the Polygenic Index Repository. Nat. Hum. Behav.). 3. To use the HRS genotype data to conduct polygenic prediction analyses for a range of social-science traits. Besides the direct interest in assessing the degree of predictive power in PGIs, we will examine how these PGIs interact with environmental factors to influence life outcomes. 4. To estimate heritability and genetic correlations for social science traits in an older population.

We will use HRS data to explore the genetic architecture of social-science outcomes. To do so, we will either use HRS data together with other datasets to identify specific genetic variants associated with these outcomes, or analyze the aggregate effect of all genetic variants in HRS alone using heritability analyses and polygenic indexes (PGIs). PGIs are summaries of a person's known genetic predisposition to a particular trait. We will use PGIs to examine the pathways underlying the relationship between genetic variants and outcomes of interest, including analyses of how genes and environment interact. We will also include HRS in future releases of the PGI Repository, an initiative that makes PGIs for a wide range of traits available in a number of datasets that may be useful to social scientists (https://www.thessgac.org/pgi-repository ). HRS was included in the first release of the Repository, and we wish to continue to update the HRS PGIs and add PGIs for new phenotypes as more data or better methodologies become available.

Research Use Statement: Our project will rely on phenotype and genotype data from the Health and Retirement Study (HRS), a nationally representative longitudinal study of the older adult population in the U.S. This is an on-going study. Data we have been using beginning in 2016 are from an approved application through dbGaP, from 15,507 HRS participants and include single nucleotide polymorphism (SNP) data on just under 2.5 million markers, imputed data on approximately 21 million DNA variants, and phenotype data on disease incidence and prevalence, functioning, biomarkers, mortality, and environmental and behavioral covariates. Our request from NIAGADS would provide us with an additional genetic sample to what we have been using, for the additional data on 3,409 participations (yielding N=18,916 total with harmonized genetic data through NIAGADS). Data usage will not create additional risk to participants. Aims of the project are to (1) Identify genetic networks and pathways that influence human aging, disease, functioning, and longevity; (2) Develop predictive models of aging-related health outcomes using information from gene networks; and (3) Examine how social and environmental conditions interact with genes within these aging-related gene networks. We will implement statistical models to test for associations between genetic variants and the same phenotype data. In moving forward with the additional samples, we will use the HRS genome-wide data to examine genetic signatures of healthspan, lifespan, and cognitive aging. Using these genetic signatures, we plan to (i) run pathway enrichment analysis to identify influential biological pathways, (ii) use them for predictive modeling of morbidity/mortality risk and cognitive aging, and (iii) incorporate information from social and behavioral data to examine GxE interactions. The overall goal of the project is to identify mechanistic gene and environment networks that contribute to aging acceleration or deceleration.

Non-Technical Summary: Aging is the largest risk factor for morbidity and mortality. Previous research using animal models or case-control studies of centenarians have suggested that variations in the pace of aging may be partially explained by genetic and genomic differences. However, few genetic regulators of human lifespan and healthspan have been identified. Furthermore, there is reason to suggest that the pace of aging may be a polygenic trait, for which multiple genes form complex networks that collectively influence aging and longevity phenotypes. These complex genetic networks may further interact with exogenous factors causing variation to arise in health outcomes under diverse environments. The goal of this project is to use advantaged statistical modeling techniques to understand how gene-gene and gene-environment interactions influence longevity and aging-related conditions.

As part of the Collaborative for Alzheimer's Disease genetics REsearch (CADRE: NIA grant U01-AG058654), we plan to analyze whole exome and whole genome sequence data generated from subjects with Alzheimer's disease (AD) and elderly normal controls. These data will be generated by the National Human Genome Institute Large-Scale Sequence Program. The goal of the planned analyses is to identify genes that have alleles that protect against or increase susceptibility to AD. We will evaluate variants detected in the sequence data for association with AD to identify protective and susceptibility alleles using the whole exome case-control data. We will also evaluate sequence data from multiplex AD families to identify variants associated with AD risk and protection, and evaluate variant co-segregation with AD. The family data will be whole genome data. The family-based data will be used to inform the cases control analysis and visa versa. We also will focus on structural variants (insertion-deletions, copy number variants, and chromosomal rearrangements). Evaluation of structural variants will involve both whole genome and whole exome data. Structural variants will be analyzed with single nucelotide variants detected and analyzed in the case-control and family-based data.

We are attempting to identify all the inherited elements that contribute to Alzheimer's disease risk. To do this we will analyze DNA sequence data from subjects with Alzheimer's disease and elderly subjects who are cognitively normal. The sequence data from these 2 groups will be compared to identify differences that contribute to the risk of developing Alzheimer's disease of that protect against Alzheimer's disease. These DNA differences can be at a single site in the genetic code, or can span multiple sites, changing the copy number of DNA sequences. Both types of genetic variants will be examined.

We have been developing more powerful statistical methods to detect common variant (CV)- or rare variant (RV)-complex trait associations and/or putative causal relationships for GWAS and DNA sequencing data. Here we propose applying our new methods, along with other suitable existing methods, to the existing ADSP sequencing data and other AD GWAS data provided by NIA, hence requesting approval for accessing the ADSP sequencing and other related GWAS/genetic data. We have the following two specific Aims: Aim1. Association testing under genetic heterogeneity: For complex traits, genetic heterogeneity, especially of RVs, is ubiquitous as well acknowledged in the literature, however there is barely any existing methodology to explicitly account for genetic heterogeneity in association analysis of RVs based on a single sample/cohort. We propose using secondary and other omic data, such as transcriptomic or metabolomic data, to stratify the given sample, then apply a weighted test to the resulting strata, explicitly accounting for genetic heterogeneity that causal RVs may be different (with varying effect sizes) across unknown and hidden subpopulations. Some preliminary analyses have confirmed power gains of the proposed approach over the standard analysis. Aim 2. Meta analysis of RV tests: Although it has been well appreciated that it is necessary to account for varying association effect sizes and directions in meta analysis of RVs for multi-ethnic cohorts, existing tests are not highly adaptive to varying association patterns across the cohorts and across the RVs, leading to power loss. We propose a highly adaptive test based on a family of SPU tests, which cover many existing meta-analysis tests as special cases. Our preliminary results demonstrated possibly substantial power gains.

We propose applying our newly developed statistical analysis methods, along with other suitable existing methods, to the existing ADSP sequencing data and other AD GWAS data to detect common or rare genetic variants associated with Alzheimer’s disease (AD). The novelty and power of our new methods are in two aspects: first, we consider and account for possible genetic heterogeneity with several subcategories of AD; second, we apply powerful meta-analysis methods to combine the association analyses across multiple subcategories of AD. The proposed research is feasible, promising and potentially significant to AD research. In addition, our proposed analyses of the existing large amount of ADSP sequencing data and other AD GWAS data with our developed new methods are novel, powerful and cost-effective.

The goal of this project is to utilize standard genetics tools and ensemble/deep learning methods to predict/classify the etiological aspects of Alzheimer's disease and other neurodegenerative diseases based on genetic data (including individual level data e.g. genotype and sequencing data and also by the use of summary statistics). Our primary phenotypes of interest include case:control status, age at onset, survival time (in terms of disease duration from diagnosis to loss to follow-up) and related biomarker data, although there may be other phenotypes of interest that are derived later based on available data.

We are attempting to identify and predict risk of Alzheimer's disease and other neurodegenerative diseases based on genetic data using standard genetic tools and advanced machine-learning methods.

Alzheimer’s disease (AD) dementia is a heterogeneous neurodegenerative disease among older adults. Early detection of AD dementia remains challenging due to heterogeneity in disease onset and progression. Our goal is to develop a genetic variants-based VAE model to predict AD dementia. We will use GWAS data collected in Health and Retirement Study (HRS). HRS has genotyped saliva DNA samples collected since 2006 at multiple time points during field visits yielding a total of 19004 unique participants. We will use the most recent version of genotype data from 2006 to 2015 for those samples that have both epigenetic and transcriptomic data available. The genotyping was performed by NIH Center for Inherited Disease Research, using the Illumina HumanOmni2.5-4v1/8v1 array, and genotyping QC analysis was performed at the University of Michigan using HumanOmni2.5-4v1 H for SNP annotation. We will use the quality and minor allele frequency (MAF) filters specified in the HRS QC report for genotypic data to filter out poor quality SNPs. We will use cognition measures collected in HRS 2016 survey to classify participants into 'Dementia' and 'Normal' using the Langa Weir Classification algorithm.We will employ two main feature selection processes: 1. Based on the association with dementia, the top 50% of associated SNPs will be selected to input to the VAE model and filter out low-frequency SNPs. 2. We will also train a VAE model with a more comprehensive list of SNPs.We will employ the model regularization by incorporating biological knowledge as constraints in the model using the gene-gene interaction network from REACTOME/ STRING. We will also evaluate the biological interpretability of latent features that are representative of input genetic variants. We will evaluate the distribution of weights of all encoded features to select the positive high and negative high features based on 2 SD above or below the mean weight. These selected features will be input for the pathway analysis to identify pathways associated with AD. The candidate genes identified can be used to develop blood-based biomarkers for early identification of AD

We will develop a genetic variants-based VAE model to predict dementia. We will employ various feature selection processes based on the complexity of data to evaluate the VAE model performance and to identify a representative list of genes. In addition, we will evaluate the biological interpretability of latent features obtained from the VAE encoder layer by extracting their decoder weights that captures the input feature contribution to the learned latent feature. This will also allow us to evaluate if the VAE model has learned novel features known to be associated with AD dementia. We will then utilize the learned features to identify the biological pathways associated with AD dementia.

We aim to uncover new genetic risk variants for Alzheimer’s disease (AD) by analysis of an integrated analysis of proteomics and genetic sequencing performed at Emory University. Results of these analyses will be used to weight analysis of whole-genome sequencing (WGS), whole-genome genotyping (WGG), and whole-exome sequencing (WES) data from dbGaP and ADSP. We plan to publish our findings, so they are shared with the scientific community.Outcomes that will be tested include: (1) clinical disease status, (2) pathologic characterization (e.g., measures of beta-amylodi, tau, etc.), and (3) cognitive decline. For sequencing data, we will perform joint calling from samples previously mapped by ADSP using PECaller using default settings. Variant annotation will be performed using Bystro and quality control will follow Wingo et al., 2017. For rare variants, we will use burden- and variance-based tests to estimate association between genetic variants and each outcome for every gene in the genome. External weights from proteomic analyses will be optionally used, as well as measures of genomic conservation for each site. For common variants, we plan to test for differences in allele frequencies using maximum likelihood tests. For all analyses, we plan to control for population structure deriving principal components from the underlying sequencing or genotyping data.

Our aim is to identify genetic variants that are associated with Alzheimer's Disease (AD) either using genomic data (from dbGap or from Emory University) or brain protein sequencing data (from Emory University) as a starting point. Each center’s data will be analyzed separately, and we will determine whether the findings are consistent among the centers. Additionally, we will use protein data from brain or cerebrospinal fluid of individuals with or without AD to guide the analysis of the genomic data to identify genetic variants that influence AD risk. Our overarching aim is to use genetic discoveries to identify mechanisms of AD pathogenesis and creation of more meaningful models of the disease.

GWAS, WES and WGS have identified many genes associated with Alzheimer’s Dementia (AD) and its related traits. However, the identified genes thus far collectively explain only a small proportion of disease heritability, suggesting that more genes remained to be identified. Moreover, there is a clear gender and ethnic disparity for AD susceptibility, but little research has been done to identify gender- and ethnic-specific variants associated with AD. Of the many challenges for deciphering AD pathology, lacking of efficient and power statistical methods for genetic association mapping and causal inference represents a major bottleneck. To tackle this challenge, we have developed a set of novel statistical and bioinformatics approaches for genetic association mapping and multi-omics causation inference in large-scale ethnicity-specific epidemiological studies. The goal of this project is to leverage the multi-omics and clinical data archived by the ADSP, ADNI, ADGC as well as other AD-related data repositories to identify novel genes and molecular markers for AD. Specifically, we will (1) validate our novel methods for identifying novel risk and protective genomic variants and multi-omics causal pathways of AD; (2) identify novel ethnicity- and gender-specific genes and molecular causal pathways of AD. We will share our results, statistical methods and computational software with the scientific community.

Although many genes have been associated with Alzheimer’s Dementia (AD), these genes altogether explain only a small fraction of disease etiology, suggesting more genes remained to be identified. Of the many challenges for deciphering AD pathology, lacking of power statistical methods represents a major bottleneck. To tackle this challenge, we have developed a set of novel statistical and bioinformatics approaches for genetic association mapping and multi-omics causation inference in large-scale ethnicity-specific epidemiological studies. The goal of this project is to leverage the rich genetic and other omic data along with clinical data archived by the ADSP, ADNI, ADGC as well as other AD-related data repositories to identify novel genes and molecular markers for AD. Such results will enhance our understanding of AD pathogenesis and may also serve as biomarkers for early diagnosis and therapeutic targets.

Alzheimer’s disease (AD) affects 5.6 million Americans over the age of 65 and exacts tremendous and increasing demands on patients, caregivers, and healthcare resources. Our current understanding of the biology and pathophysiology of AD is still limited, hindering advances in the development of therapeutic and preventive strategies. Existing genetic studies of AD have some success but these explain only a fraction of the overall disease risk, suggesting opportunities for additional discoveries. The proposed project will leverage existing neuroimaging and genetic data resources from the UK Biobank, the Alzheimer’s Disease Sequencing Project (ADSP), the Alzheimer’s Disease Neuroimaging Initiative (ADNI), and the Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE) consortium, and will be conducted by a multidisciplinary team of investigators. We will derive AD endophenotypes from neuroimaging data in the UK Biobank using deep learning (DL). We will identify novel genetic loci associated with DL-derived imaging endophenotypes and optimize the co-heritability of these endophenotypes with AD-related phenotypes using UK Biobank genetic data. We will leverage resources and collaborations with AD Consortia and the power of DL-derived neuroimaging endophenotypes to identify novel genes for Alzheimer’s Disease and AD-related traits. Also, we will develop DL-based neuroimaging harmonization and imputation methods and distribute implementation software to the research community. We expect to discover new genes relevant to AD which may leads to understanding of molecular basis of AD and potential new treatment.

Alzheimer’s disease (AD) exacts a tremendous burden on patients, caregivers, and healthcare resources. Our current understanding of the biology of AD is still limited, hindering advances in the development of treatment and prevention. Existing genetic studies of AD have some success but more studies are needed. The proposed project will leverage existing neuroimaging and genetic data resources from the UK Biobank, the Alzheimer’s Disease Sequencing Project (ADSP) and other consortia and will be conducted by a multidisciplinary team of investigators. We will derive new AD relevant intermediate phenotypes from neuroimaging data using deep learning (DL), an AI approach. We will identify novel genetic loci associated with these phenotypes. Also, we will develop imaging harmonization and imputation methods and distribute implementation software to the research community. We expect to discover new genes relevant to AD which may leads to understanding of molecular basis of AD and potential new treatment.