Research Projects Directory

Genetic association studies - Zhi's group

Our overarching goal is to characterize the genetic architecture of complex traits—such as Alzheimer’s disease, anthropometric measures (e.g. height), and metabolic biomarkers—by combining standard genome-wide association studies (GWAS) with identity-by-descent (IBD) mapping. GWAS will uncover common variant associations across hundreds…

Scientific Questions Being Studied

Our overarching goal is to characterize the genetic architecture of complex traits—such as Alzheimer’s disease, anthropometric measures (e.g. height), and metabolic biomarkers—by combining standard genome-wide association studies (GWAS) with identity-by-descent (IBD) mapping. GWAS will uncover common variant associations across hundreds of phenotypes, while IBD analysis will leverage recent shared ancestry to boost power for rare and low-frequency variants that are typically underpowered in conventional analyses. We will infer IBD segments at biobank scale, build genome-wide IBD matrices, and use them in mixed-model association tests to localize both common and rare variant signals. By integrating GWAS and IBD approaches, we aim to illuminate novel genetic contributors to human health and disease—knowledge that could inform risk prediction, identify therapeutic targets, and advance precision medicine for diverse populations.

Project Purpose(s)

Methods Development
Ancestry

Scientific Approaches

We will use genotype and phenotype data from the All of Us dataset. Standard GWAS analyses will be performed using efficient software such as PLINK and SAIGE to identify common variant associations. To capture rare variant associations, we will infer identity-by-descent (IBD) segments genome-wide using scalable tools like hap-IBD, Refined IBD, and FiMAP. The inferred IBD segments will be summarized into genome-wide IBD matrices, enabling mixed-model analyses with tools such as GMMAT or HiFiMAP to detect associations driven by recent ancestry and rare variants. Results will be visualized and interpreted using standard bioinformatics pipelines.

Anticipated Findings

We anticipate identifying novel genetic loci associated with common and rare variation underlying complex traits, such as Alzheimer's disease and anthropometric phenotypes. GWAS analyses will uncover associations driven by common variants, while IBD-based methods will uniquely enhance our ability to detect rare variants that GWAS alone often misses. These findings will improve understanding of the genetic architecture of diverse traits, refine risk prediction models, and potentially highlight therapeutic targets, significantly advancing precision medicine and genomic research in diverse populations.

Demographic Categories of Interest

This study will not center on underrepresented populations.

Data Set Used

Controlled Tier

Research Team

Owner:

Bohong Guo - Graduate Trainee, University of Texas Health Science Center, Houston

DS pregnancy and AD

We want to understand whether females carrying a Down syndrome pregnancy are at greater risk of developing Alzheimer's Disease later in life compared to females who do not carry a Down syndrome pregnancy. Early research from the 1990s suggested an…

Scientific Questions Being Studied

We want to understand whether females carrying a Down syndrome pregnancy are at greater risk of developing Alzheimer's Disease later in life compared to females who do not carry a Down syndrome pregnancy. Early research from the 1990s suggested an increased risk in females if they had a Down syndrome pregnancy before the age of 35, but not after.

Project Purpose(s)

Disease Focused Research (Alzheimer Disease)

Scientific Approaches

We want to carry out an association study by comparing cognitive function and potential Alzheimer's Disease prognosis between women with pregnancies in general, and pregnancies with problems, including miscarriage, as well as pregnancies with Down syndrome.

Anticipated Findings

We expect to validate previous findings that women with Down syndrome pregnancies have an increased risk of Alzheimer's Disease.

Demographic Categories of Interest

This study will not center on underrepresented populations.

Data Set Used

Controlled Tier

Research Team

Owner:

Kilian Kleemann - Graduate Trainee, University Hospital Bonn

cardiovascular_phenotypes

Are there novel variants associated with cardiovascular disease phenotypes we can pick up using a larger, more diverse dataset? Do these variants overlap with any cognitive phenotypes like Alzheimer's Disease, Parkinson's Disease, etc?

Scientific Questions Being Studied

Are there novel variants associated with cardiovascular disease phenotypes we can pick up using a larger, more diverse dataset? Do these variants overlap with any cognitive phenotypes like Alzheimer's Disease, Parkinson's Disease, etc?

Project Purpose(s)

Disease Focused Research (Cardiovascular diseases)

Scientific Approaches

I plan to use genome-wide association studies (GWAS) vis PLINK and SAIGE. I will create cohorts based on the appearance of certain ICD10-based codes in patients' EHR for cardiovascular phenotypes.

Anticipated Findings

We anticipate finding novel variants linked to cardiac disease and that some may overlap with cognitive phenotypes as well.

Demographic Categories of Interest

This study will not center on underrepresented populations.

Data Set Used

Controlled Tier

Research Team

Owner:

Scott Dudek - Project Personnel, University of Pennsylvania
Anni Moore - Graduate Trainee, University of Pennsylvania

Collaborators:

Yuki Bradford - Project Personnel, University of Pennsylvania

AoU V8 Linking Neurodevelopmental and Neurodegenerative Disorders

The purpose of this study is to examine the relationship between neurodevelopment (including conditions like autism and ADHD), genetic markers of such conditions, and neurodegeneration (including dementia, Alzheimer’s, and Parkinson’s). We will examine the role of several intervenable targets between…

Scientific Questions Being Studied

The purpose of this study is to examine the relationship between neurodevelopment (including conditions like autism and ADHD), genetic markers of such conditions, and neurodegeneration (including dementia, Alzheimer’s, and Parkinson’s). We will examine the role of several intervenable targets between neurodevelopmental and neurodegeneration. These lifestyle factors are known to impact neurodegeneration in the general population, but their role in people with neurodevelopmental conditions is unknown. The purpose of examining these factors is to examine if the risk of cognitive decline may be mitigated via improvements to these targets. These include cardiometabolic health (including hypertension, obesity, and diabetes), cognitive reserve (including higher education levels), physical activity, and accelerated biological aging.

Project Purpose(s)

Disease Focused Research (Autism, intellectual disability, and attention deficit/hyperactivity disorder)
Population Health
Ancestry

Scientific Approaches

We will compare rates of neurodegenerative disorders across people with genetic liability for autism, ADHD, and ID, identified via polygenic risk scores, which will be expressed as a z-score based on the mean/standard deviation. We will examine data across the life course for evidence that genetic risk for neurodevelopmental conditions is linked to neurodegeneration and we will estimate the genetic correlation between the two using cross-trait linkage disequilibrium score regression. We will then examine modifiable factors that may play a role in reducing risk, creating indicators from survey questions about lifestyle. We will use DNA methylation data to construct epigenetic clocks to examine accelerated aging. We will similarly examine the impact of other modifiable factors including cardiometabolic health, cognitive reserve (measured via employment and educational attainment), and self-reported physical activity levels.

Anticipated Findings

Preliminary studies have suggested neurodegeneration may be more pronounced in people with neurodevelopmental conditions. The ultimate purpose of this research is to increase understanding of neurodegeneration risk in people with neurodevelopmental conditions, with an eye towards modifiable factors. We will also examine the genetic link between genes associated with neurodevelopmental conditions, which may illuminate causal pathways to neurodegeneration. Using high-resolution phenotyping available in All of Us uniquely allows the examination of early-onset degeneration. This is important, as early-onset degeneration may be a target for intervention. Other modifiable lifestyle factors, such as physical activity and cardiovascular health, may also be important targets for intervention. These explorations hope to improve public health. Overall, this study aims to expand our understanding of dementia in neurodivergent people, with the ultimate goal of improving health and well-being.

Demographic Categories of Interest

Data Set Used

Controlled Tier

Research Team

Owner:

Jessica Rast - Research Associate, Drexel University

Collaborators:

Jacob Bergstedt - Research Fellow, Drexel University

Alzheimer & Exposome

create a workspace for the Alzheimer’s diagnosed group to see what we can glean about their metabolic health status from EHR and compare with age sex race ethnicity matched healthy.

Scientific Questions Being Studied

create a workspace for the Alzheimer’s diagnosed group to see what we can glean about their metabolic health status from EHR and compare with age sex race ethnicity matched healthy.

Project Purpose(s)

Educational

Scientific Approaches

PCA, OPLS-DA, etc. create a workspace for the Alzheimer’s diagnosed group to see what we can glean about their metabolic health status from EHR and compare with age sex race ethnicity matched healthy.

Anticipated Findings

unknown at the moment. it will contributeto if we're seeing association between alzheimer's metabolic health status from exposome.

Demographic Categories of Interest

This study will not center on underrepresented populations.

Data Set Used

Registered Tier

Research Team

Owner:

Yuheng Che - Graduate Trainee, University of North Carolina, Chapel Hill

AD and CVD Associations

I hope to further understand the differing associations between Alzheimer's disease (AD) and cardiovascular disease (CVD). Growing evidence supports a strong relationship between heart and brain health, and it is important to understand which forms of CVD are more closely…

Scientific Questions Being Studied

I hope to further understand the differing associations between Alzheimer's disease (AD) and cardiovascular disease (CVD). Growing evidence supports a strong relationship between heart and brain health, and it is important to understand which forms of CVD are more closely associated with AD to gain insight into the mechanisms behind dementia, as well as CVD patients who may be at heightened risk for dementia.

Project Purpose(s)

Disease Focused Research (Alzheimer's Disease, Cardiovascular Disease)

Scientific Approaches

I will utilize the All of Us dataset to validate the results I found when performing this analysis in the UK Biobank. I will calculate adjusted odds ratios (OR) for AD, given the diagnosis of 11 different forms of CVD. This model will be adjusted for well-established covariates between CVD and AD. The analysis will be performed using RStudio.

Anticipated Findings

I expect that the findings of this study will show significant relationships between most of the CVD subtypes and AD. I hope to find similar results to what I found in the UK Biobank. However, if there are differences, these could show me how the relationship between AD and CVD varies across different populations. The UK Biobank is an overwhelmingly white British cohort, while the All of Us data shows how these relationships may differ in a more diverse American population.

Demographic Categories of Interest

This study will not center on underrepresented populations.

Data Set Used

Controlled Tier

Research Team

Owner:

Aili Toyli - Undergraduate Student, Michigan Technological University

Association of an ABCA7 tandem repeat with Alzheimer's Disease

We plan to explore the involvement of a genetic variant in a gene called ABCA7 with Alzheimer's Disease. This variant is a "tandem repeat", a kind of variant that is often underexplored in genetic studies but that has a high…

Scientific Questions Being Studied

We plan to explore the involvement of a genetic variant in a gene called ABCA7 with Alzheimer's Disease. This variant is a "tandem repeat", a kind of variant that is often underexplored in genetic studies but that has a high mutation rate, and many examples of which have been associated with human health. This variant has been linked with Alzheimer's Disease before, but we hope to validate that finding in the AllOfUs cohort and further explore the data. Exploring this topic will help researchers understand the cause of Alzheimer's Disease and ultimately could lead to treatments or cures.

Project Purpose(s)

Disease Focused Research (Alzheimer's Disease)
Ancestry

Scientific Approaches

We will analyze the long-read sequencing (LRS) data files in order to genotype the tandem repeat variant and associate the length of this tandem repeat with the Alzheimer's Disease "proxy phenotype" (i.e., whether a given individual or any of their first-degree relatives have had dementia).

Anticipated Findings

We anticipate that the length of the VNTR will be associated with Alzheimer's Disease. Reproducing this finding in this dataset will help strengthen the previously identified association, and we will leverage the data to further understand what length threshold for the variant constitutes a "pathogenic" copy of the variant.

Demographic Categories of Interest

Data Set Used

Controlled Tier

Research Team

Owner:

Andrés Peña-Tauber - Undergraduate Student, Stanford University

Duplicate of Duplicate of ALDH2 HLP

Alcohol consumption is a risk factor for many chronic diseases, including some cancers, type 2 diabetes, and Alzheimer’s disease. Individuals with a specific variant of the aldehyde dehydrogenase 2 gene (ALDH2*2) are at higher risk of many of these diseases.…

Scientific Questions Being Studied

Alcohol consumption is a risk factor for many chronic diseases, including some cancers, type 2 diabetes, and Alzheimer’s disease. Individuals with a specific variant of the aldehyde dehydrogenase 2 gene (ALDH2*2) are at higher risk of many of these diseases. Given that ALDH2*2 is the most common single genetic variation in humans and that more than half of all American adults drink alcohol, an opportunity is present for targeted chronic disease risk reduction in a large number of Americans. However, in order to design effective public health strategies, such as targeted intervention programs, a better understanding of current alcohol consumption behaviors and associated factors, overall and stratified by ALDH2 genotype, is needed. This study aims to characterize the alcohol consumption behaviors among participants in the All of Us Research Program and examine factors that may be related to the behaviors, overall and by ALDH2 genotype.

Project Purpose(s)

Population Health
Social / Behavioral
Ancestry

Scientific Approaches

We will analyze data from the All of Us Research Program database regarding alcohol consumption behaviors, ALDH2 genotype (rs671), demographics, personal and family health history, socioeconomic factors, lifestyle factors, and social determinants of health. All participants with informative data for rs671 will be included in the study. These data sets will be from surveys, physical measurements, and the genomic data set. Statistical analyses will be done using R or python. We will examine relationships between these factors and alcohol consumption using Fisher’s exact test (categorical variables) and the Kruskal-Wallis test (continuous variables), overall and stratified by ALDH2 genotype and potentially other factors, for example, demographics.

Anticipated Findings

We hypothesize that alcohol consumption behaviors will be associated with factors including demographics, personal and family health history, socioeconomic factors, lifestyle factors, and social determinants of health, with ALDH2 genotype and potentially other factors. While a limited number of U.S. studies among university students have shown that ALDH2*2 homozygotes tend to avoid alcohol, many ALDH2 heterozygotes do consume alcohol, albeit at lower levels. There have been no studies examining alcohol consumption behaviors by ALDH2 genotype conducted outside the university setting in the U.S. The All of Us Research Program presents a valuable source of data from study participants across the U.S. which would enable the study of alcohol consumption behaviors in the context of genomics.

Demographic Categories of Interest

Data Set Used

Controlled Tier

Research Team

Owner:

Manxi Yang - Graduate Trainee, University of California, Irvine
Hannah Park - Other, University of California, Irvine
Danielle Forman - Other, University of California, Irvine

Collaborators:

Jacqueline Kim - Other, University of California, Irvine
Hester Nguyen - Undergraduate Student, University of California, Irvine
Sara Ghaem-Maghami - Undergraduate Student, University of California, Los Angeles

Investigation of Genes implicated in Alzheimers and their connection to cLQTs

Recent studies have shown that patients with dementia have a higher risk for cLQTs (Danese). Alzheimers is a type of dementia that affects millions of people world wide and its disease pathway is not fully understood. The goal of this…

Scientific Questions Being Studied

Recent studies have shown that patients with dementia have a higher risk for cLQTs (Danese). Alzheimers is a type of dementia that affects millions of people world wide and its disease pathway is not fully understood. The goal of this study will be to determine if there are any genes that may be linking Alzheimers and lQTs. We hope that by identifying genes that link these diseases, doctors will be better equipped to treat patients carrying these mutations through early detection and treatment.

Danese A, Federico A, Martini A, et al. QTc Prolongation in Patients with Dementia and Mild Cognitive Impairment: Neuropsychological and Brain Imaging Correlations. Journal of Alzheimer’s Disease. 2019;72(4):1241-1249. doi:10.3233/JAD-190632

Project Purpose(s)

Disease Focused Research (Long QT Syndrome)
Methods Development
Ancestry

Scientific Approaches

The goal is to use multiple PheWAS's that each focus on one gene. These genes include (APP, APOE, Psen1, Psen2 ...). The information from this PheWAS will be used solely to validate our idea that Alzheimers and lQTs are linked. If we find genes that show a possible linkage then we will move into cell work and use gene Knockouts and Knockdowns to better understand that affects of the gene on lQTc. We will do this because Alzhiemers and lQTs share a lot of similar risk factors and the only ways to ensure that these genes actually play a role in lQTc is by working in cells. We will also be looking to identify and understand the signaling pathway that the implicated gene is part of. This will likely be done using RNA sequencing and its results will not only yield the pathway, but possible drug targets

Anticipated Findings

This database is solely being used as a proof of concept. We suspect that we will find a genetic linkage between the two diseases and will be able to transition the experiment into the wet lab setting. The overall hope of the experiment is to establish the genetic linkage so that doctors can be more vigilant for Alzheimers or lQTs in patients who carry the gene and currently have one of the diseases. The other goal is to identify a new pathways which will hopefully present novel drug targets that can be used to treat lQTs.

Demographic Categories of Interest

Data Set Used

Controlled Tier

Research Team

Owner:

Armann Gudmundsson - Undergraduate Student, Vanderbilt University

AD and diabetes

Recently, GWAS and proteomics data have been used to examine the association between genotypes and protein abundance that has promoted the emergence of data analyses, which facilitate the combination of multidimensional data and integrate protein expression data with GWAS findings…

Scientific Questions Being Studied

Recently, GWAS and proteomics data have been used to examine the association between genotypes and protein abundance that has promoted the emergence of data analyses, which facilitate the combination of multidimensional data and integrate protein expression data with GWAS findings in Alzheimer's disease (AD) and diabetic pathogenesis.

Project Purpose(s)

Disease Focused Research (dementia and diabetes)
Population Health
Social / Behavioral
Control Set
Ancestry

Scientific Approaches

The data from GWAS will be analyzed by following steps:
1. Performing GWAS using data from the All of Us research program to identify potential genes related to AD and diabetes.
2. Conducting protein quantitative trait locus (QTL) analysis of 191,440 proteins in serum or plasma to detect associations between AD-associated genotypes and protein abundance.

Anticipated Findings

Overall, this grant proposal seeks to advance our understanding of the genetic basis and protein biomarkers of AD and diabetes and identify novel therapeutic targets for the disease.

Demographic Categories of Interest

Data Set Used

Controlled Tier

Research Team

Owner:

Chun (Cathy) Xu - Mid-career Tenured Researcher, University of Texas, Rio Grande Valley

Collaborators:

Kesheng Wang - Teacher/Instructor/Professor, University of South Carolina
Belinda Alvarado - Graduate Trainee, University of Texas, Rio Grande Valley
Bassent Abdelbary - Other, University of Texas, Rio Grande Valley

Duplicate

Mitochondrial DNA mutations have harmful consequences in 1 in 5000 people such as Alzheimer’s and Parkinson’s, common age-related diseases. With the overall goal to prevent and repair damage from mitochondrial DNA mutations, MitoSENS’s focus is to develop ways to engineer…

Scientific Questions Being Studied

Mitochondrial DNA mutations have harmful consequences in 1 in 5000 people such as Alzheimer’s and Parkinson’s, common age-related diseases. With the overall goal to prevent and repair damage from mitochondrial DNA mutations, MitoSENS’s focus is to develop ways to engineer cells to make these mutations harmless to aging cells. To better understand the relevance of mtDNA mutations in aging, my research project further involves analyzing large genomic datasets such as the All of US dataset for variance across the mitochondrial genes. The results from this analysis will help us define patterns in relation to age and disease subsets.

Project Purpose(s)

Disease Focused Research (Alzheimer's disease, mitochondrial disorders, Parkinson's)
Ancestry

Scientific Approaches

I plan to begin with a small, diverse sample size (n=100) with focus on age. I will use data analysis skills through Python programming (Pandas) and variant analysis tools (ie. BCFtools, VCFtools) to answer this question. I may utilize mitochondrial-specific tools but I will update this section if need be.

Anticipated Findings

The results from this analysis will help us define patterns in relation to age and disease subsets. My findings will contribute to regenerative medicine, which is focused on developing new treatments to restore function lost in tissues and organs due to aging, disease, damage or defects. Under standing the specific mutations that accumulate as we age can help us come one step closer to coming up with new therapies for age-related, or broadly mitochondrial, disorders.

Demographic Categories of Interest

This study will not center on underrepresented populations.

Data Set Used

Controlled Tier

Research Team

Owner:

Nitya Jain - Research Assistant, Lifespan Research Institute

ABCA1 and APOE genetic interaction in Alzheimer's Disease

We are hoping to confirm a hypothesis we have developed using other large-scale genetic data. The hypothesis is that two genes independently implicated in Alzheimer's Disease, ABCA1 and APOE, interact (that is, are dependent on each other) to change disease…

Scientific Questions Being Studied

We are hoping to confirm a hypothesis we have developed using other large-scale genetic data. The hypothesis is that two genes independently implicated in Alzheimer's Disease, ABCA1 and APOE, interact (that is, are dependent on each other) to change disease risk. Answering this question could help understand the mechanisms of Alzheimer's Disease and inform therapeutic strategies that target those two genes.

Project Purpose(s)

Disease Focused Research (Alzheimer's Disease)
Ancestry

Scientific Approaches

We will use whole-genome sequencing to assess for rare variants in ABCA1 and analyze them against Alzheimer's Disease risk using the "proxy" phenotype (whether each individual or a direct relative has been diagnosed with the disease). We will then use a regression approach, called Cox hazards regression, that will tell us whether the interaction between variants on ABCA1 that are predicted to affect protein function and APOE genotype are correlated with Alzheimer's Disease. We will study this question both in European and African ancestry populations.

Anticipated Findings

The anticipated finding is that ABCA1 and APOE interact with each other genetically to change Alzheimer's Disease risk. This would contribute to the understanding of these two genes in the disease, which could help the development of therapeutics targeting them.

Demographic Categories of Interest

Data Set Used

Controlled Tier

Research Team

Owner:

Andrés Peña-Tauber - Undergraduate Student, Stanford University

Investigating SNPs in Mitochondrial and Mitophagy genes in AD v8

Scientific Questions: 1. Are there specific SNP variants in mitochondrial biogenesis and mitophagy genes that are significantly associated with Alzheimer's disease (AD)? 2. Do these SNP variants show differential frequencies in AD vs. non-AD populations? 3. How do these SNPs…

Scientific Questions Being Studied

Scientific Questions:
1. Are there specific SNP variants in mitochondrial biogenesis and mitophagy genes that are significantly associated with Alzheimer's disease (AD)?
2. Do these SNP variants show differential frequencies in AD vs. non-AD populations?
3. How do these SNPs affect gene expression and mitochondrial function in AD neurons?

Why is this important:
While amyloid and tau pathologies have dominated AD research, mitochondrial dysfunction is an important but understudied contributor. This study integrates genomics and computational biology to uncover how genetic variants in mitochondrial pathways influence AD progression.

Project Purpose(s)

Disease Focused Research (Alzheimer's disease)
Ancestry

Scientific Approaches

The study will utilize a computational genomics approach to identify SNP variants in mitochondrial biogenesis and mitophagy genes associated with AD. the study will follow a case-control design, leveraging genomic and electronic health records (EHR) data from the All of Us Database to investigate genetic contributions to mitochondrial dysfunction in AD. We will perform whole-genome association studies to compare SNP frequencies in AD vs controls (age-matched non-AD individuals). We will also determine if these SNPs occur in the coding or regulatory regions (this will help determine if the SNPs affect gene expression or protein function of these mitochondrial and mitophagy genes in AD). We will use the EHR data to link SNPs to AD phenotypes.

Anticipated Findings

1. SNPs in mitochondrial biogenesis and mitophagy genes will show a higher frequency in AD compared to non-AD.
2. Discovery of regulatory or exonic SNPs that reduce mitochondrial biogenesis and mitophagy gene expression or protein function, impairing mitochondrial function in AD.
3. Some SNPs will correlate with AD phenotypes, e.g. PPARGC1A and TFAM SNPs (linked to faster hippocampal atrophy, worsening memory loss).

Demographic Categories of Interest

This study will not center on underrepresented populations.

Data Set Used

Controlled Tier

Research Team

Owner:

Nobuki Hida - Graduate Trainee, University of North Dakota
Junguk Hur - Mid-career Tenured Researcher, University of North Dakota

Duplicate of AD and diabetes

Recently, GWAS and proteomics data have been used to examine the association between genotypes and protein abundance that has promoted the emergence of data analyses, which facilitate the combination of multidimensional data and integrate protein expression data with GWAS findings…

Scientific Questions Being Studied

Recently, GWAS and proteomics data have been used to examine the association between genotypes and protein abundance that has promoted the emergence of data analyses, which facilitate the combination of multidimensional data and integrate protein expression data with GWAS findings in Alzheimer's disease (AD) and diabetic pathogenesis.

Project Purpose(s)

Disease Focused Research (dementia and diabetes)
Population Health
Social / Behavioral
Control Set
Ancestry

Scientific Approaches

The data from GWAS will be analyzed by following steps:
1. Performing GWAS using data from the All of Us research program to identify potential genes related to AD and diabetes.
2. Conducting protein quantitative trait locus (QTL) analysis of 191,440 proteins in serum or plasma to detect associations between AD-associated genotypes and protein abundance.

Anticipated Findings

Overall, this grant proposal seeks to advance our understanding of the genetic basis and protein biomarkers of AD and diabetes and identify novel therapeutic targets for the disease.

Demographic Categories of Interest

Data Set Used

Controlled Tier

Research Team

Owner:

Chun (Cathy) Xu - Mid-career Tenured Researcher, University of Texas, Rio Grande Valley

Collaborators:

Kesheng Wang - Teacher/Instructor/Professor, University of South Carolina
Belinda Alvarado - Graduate Trainee, University of Texas, Rio Grande Valley
Bassent Abdelbary - Other, University of Texas, Rio Grande Valley

ADRD & Black Adults Nativity

To assess the sociocultural factors within the distinct subgroups constituting the prevalence of Alzheimer's Disease and Related Dementias (ADRD) among U.S. Black population. We investigated the prevalence and distribution of ADRD within the US Black population, stratified by nativity (US-born…

Scientific Questions Being Studied

To assess the sociocultural factors within the distinct subgroups constituting the prevalence of Alzheimer's Disease and Related Dementias (ADRD) among U.S. Black population. We investigated the prevalence and distribution of ADRD within the US Black population, stratified by nativity (US-born vs. non-US-born). Our findings provide valuable insights into the complex interplay of factors influencing ADRD risk and outcomes within this population.

Project Purpose(s)

Disease Focused Research (ADRD)
Social / Behavioral
Ethical, Legal, and Social Implications (ELSI)

Scientific Approaches

The data source for this study will be the All of Us Research Workbench. We have created a workspace entitled “ADRD & Black African Adults Nativity” on the All of Us Researcher Workbench. and used dataset builder to create datasets, and export the data to Jupyter Notebooks for analysis using R and Python V3.0.

We created a workspace named "ADRD & Black Adults Nativity" within the AoU Researcher Workbench using the controlled tier data. We constructed two distinct cohorts based on demographic information from the AoU Basic Survey: one consisting of US-born Black individuals the other of non-US-born Black individuals. We extracted ADRD diagnosis data from participants' EHR, utilizing International Classification of Diseases (ICD) codes (ICD-9 and ICD-10) specific to ADRD and mild cognitive impairment (MCI). Additionally, we incorporated standard concept groups derived from ICD coding and classification.

Anticipated Findings

The prevalence of Alzheimer's Disease and Related Dementias (ADRD) among native-born and non-native-born US Black individuals would be different by age, race, and socioeconomic/environmental factors including income, education, and neighborhood characteristics

Demographic Categories of Interest

Data Set Used

Controlled Tier

Research Team

Owner:

Yijiong Yang - Research Fellow, Florida State University

Neuroscience_Genetics

“Protective rare variants, epistasis and repeat expansions in Neuroscience” The Neuroscience Genetics team at Eli Lilly leverages human genetic data to identify and validate novel molecular pathways and therapeutic targets through variant-gene-disease associations related to neurodegenerative phenotypic outcomes. Our primary…

Scientific Questions Being Studied

“Protective rare variants, epistasis and repeat expansions in Neuroscience”
The Neuroscience Genetics team at Eli Lilly leverages human genetic data to identify and validate novel molecular pathways and therapeutic targets through variant-gene-disease associations related to neurodegenerative phenotypic outcomes. Our primary focus is on uncovering genetic associations for Alzheimer's disease (AD), amyotrophic lateral sclerosis (ALS), Parkinson's disease (PD), chronic pain and other rare neurodegenerative diseases. Utilizing advanced bioinformatics tools, we perform comprehensive analyses of genetic data to identify disease loci or gene-associated traits. Beyond genome-wide association studies (GWAS) approaches and non-coding common variants identification, we plan to expand our scope to investigate the impact of rare variants at the individual level, notably protective variants that have shown efficacy in other neurological and pain disorders.

Project Purpose(s)

Disease Focused Research (AD, PD, ALS, chronic pain, rare neurodegenerative diseases)
Methods Development
Control Set
Ancestry
Commercial

Scientific Approaches

Identify relevant groups of patients with neurodegenerative disorders.
Perform genetic analysis of these participants to identify primary genetic associations and protective or risk genetic variants that modulate age at onset, penetrance and rate of progression. This will include both in-depth analysis of specific genes and unbiased genome-wide approaches using whole genome sequencing/array genotyping data.
Identify clinical and environmental factors that modulate or are associated with these genetic associations, including chronic pain conditions or other co-morbidities such as diabetes, psychiatry or autoimmune disorders. Cases will be designed based on encoded combinations of phenotypic endpoints and aggregated combinations derived from primary diagnoses, comorbidities, biomarkers, presence of genetic mutations, pathogenic repeats and questionnaires.

Anticipated Findings

With the current research plan we will investigate the genetic architecture of neuroscience indications with the objective to identify novel therapeutic targets and subsequent development of genetic medicines using advanced analysis methods. In addition to known genetic associations and causal mutations, we will focus on methods to identify protective rare variants and mutations that modulate the penetrance and expressivity of variants in disease-causing genes. Mutations leading to odd ratios far below one may serve as starting point for novel therapeutic interventions that protect from disease onset or slow the rate of decline.
Ultimately these efforts will lead to the identification of targets that delay the onset of disease, slow the rate of decline, and treat comorbidities. These specific genetic factors may become attractive targets and genetic editing options for the development of novel treatments for patients.

Demographic Categories of Interest

This study will not center on underrepresented populations.

Data Set Used

Controlled Tier

Research Team

Owner:

Achim Kless - Senior Researcher, Eli Lilly and Company

Collaborators:

YoSon Park - Senior Researcher, Eli Lilly and Company
Elisa Molinari - Senior Researcher, Eli Lilly and Company
Jocelyn Quistrebert - Other, Eli Lilly and Company
Aidan Nickerson - Other, Eli Lilly and Company

Alzheimer's disease prediction algorithm

The goal of this work is to assess prediction accuracy of a mathematical algorithm for early diagnosis of Alzheimer's disease (AD). Specifically, the main aim is to determine which combination of biomarkers and risk factors results in the best diagnostic…

Scientific Questions Being Studied

The goal of this work is to assess prediction accuracy of a mathematical algorithm for early diagnosis of Alzheimer's disease (AD). Specifically, the main aim is to determine which combination of biomarkers and risk factors results in the best diagnostic tool. The goal is to enable earlier and more precise AD diagnosis than is currently possible and thus reach more patients in early stages and thereby extend their treatment time and enhance the treatment benefits.

Project Purpose(s)

Disease Focused Research (Alzheimer's disease)
Methods Development

Scientific Approaches

The AD diagnostic algorithm studied here, is a parametric risk function with two parts: 1) ‘Base curve’ describing disease probability at a certain age for men and women; 2) Relative risks associated with family history of AD, biomarkers and physical measurements. To avoid overfitting, model parameters are based on reported estimates from previous studies that did not include the All of US dataset. The analysis will be based on the largest possible set of individuals with relevant information. Prediction accuracy will be determined from how well the algorithm can differentiate between AD patients and controls in the All of US dataset. It will also be examined if the algorithm can reliably compute the risk of disease onset within a certain time frame. The analysis will compare predictive performance for different combinations of input variables

Anticipated Findings

We expect to attain information about the prediction accuracy of the algorithm as well as what specific combination of input variables yields the best diagnostic tool for AD.

Demographic Categories of Interest

This study will not center on underrepresented populations.

Data Set Used

Registered Tier

Research Team

Owner:

Sigríður Rós Einarsdóttir - Graduate Trainee, University of Iceland

Collaborators:

Stefania Benonisdottir - Research Associate, University of Iceland

Alzheimer EDA

I plan to study how well the current CMS Hierarchical Condition Category (HCC) risk score predicts healthcare needs in older adults with Alzheimer’s disease. This question is important because risk scores are used to determine funding and support for patients,…

Scientific Questions Being Studied

I plan to study how well the current CMS Hierarchical Condition Category (HCC) risk score predicts healthcare needs in older adults with Alzheimer’s disease. This question is important because risk scores are used to determine funding and support for patients, but they may not fully reflect the needs of people with cognitive decline. I want to explore whether certain health patterns or conditions are being missed in this group, and if so, how the risk score might be improved. This research could help improve care planning and policy decisions for older adults with memory-related conditions.

Project Purpose(s)

Disease Focused Research (Alzheimer's)
Educational

Scientific Approaches

I will use data from the All of Us Research Program, focusing on older adults diagnosed with Alzheimer’s disease. I plan to analyze their medical conditions, medication use, lab results, and clinical visits using the OMOP Common Data Model. My approach includes identifying patterns and comparing risk scores calculated using the CMS-HCC model. I will use descriptive statistics and basic predictive modeling techniques to examine how well current risk scores reflect the healthcare needs of these patients. Tools I plan to use include R, Python, and Jupyter Notebooks in the Researcher Workbench. This study is for educational purposes only as part of my graduate research project and is not intended for clinical or commercial use.

Anticipated Findings

I anticipate finding that the current CMS-HCC risk score may not fully capture the healthcare needs of older adults with Alzheimer’s disease. The model may overlook important clinical patterns related to cognitive decline, leading to underestimation of risk. By identifying these gaps, this study could suggest ways to improve the accuracy of risk scores for this population. These findings would add to the scientific understanding of how risk models work for patients with complex conditions like Alzheimer’s. This could help researchers, policymakers, and healthcare providers better support older adults with memory-related illnesses. This project is for educational purposes only and will not be used for clinical decisions.

Demographic Categories of Interest

This study will not center on underrepresented populations.

Data Set Used

Controlled Tier

Research Team

Owner:

Jaideep Solanki - Graduate Trainee, Harrisburg University of Science and Technology

indigenous cohort Alzheimer's analysis

Livingston-style health analysis of the Indigenous cohort, looking at controllable risk factors for Alzheimer's

Scientific Questions Being Studied

Livingston-style health analysis of the Indigenous cohort, looking at controllable risk factors for Alzheimer's

Project Purpose(s)

Population Health

Scientific Approaches

Logistic regression and similar forms of statistical analysis (subject to change depending on data format)

Anticipated Findings

Looking at possibly controllable risk factors for Alzheimer's in the Indigenous population, may help tribal officials make public health policy and spread community knowledge

Demographic Categories of Interest

Data Set Used

Controlled Tier

Research Team

Owner:

Carina Campbell - Undergraduate Student, University of Wisconsin, Madison

Collaborators:

Kyle Conniff - Research Fellow, University of Wisconsin, Madison

Complex traits, GWA and protein

Recently, GWAS and proteomics data have been used to examine the association between genotypes and protein abundance that has promoted the emergence of data analyses, which facilitate the combination of multidimensional data and integrate protein expression data with GWAS findings…

Scientific Questions Being Studied

Recently, GWAS and proteomics data have been used to examine the association between genotypes and protein abundance that has promoted the emergence of data analyses, which facilitate the combination of multidimensional data and integrate protein expression data with GWAS findings in Alzheimer's disease (AD) pathogenesis. All of Us Research Program offers an opportunity for researcher to identify AD associated biomarkers.
The aims of the current research proposal are to conduct a comprehensive analysis of GWAS and proteomics data to unravel the genetic underpinnings of AD using the All of US Researcher Workbench. Our analysis will focus on identifying rare variants with large effect sizes and common variants with smaller effect sizes that may contribute to AD susceptibility among different ethnic groups, such as Hispanic population. This integrative approach will provide insights into the molecular mechanisms underlying AD and may reveal novel therapeutic targets for the disease.

Project Purpose(s)

Disease Focused Research (Alzheimer's disease)
Population Health
Control Set
Ancestry

Scientific Approaches

The data from GWAS will be analyzed by following steps:
1. Performing GWAS using data from the All of Us research program (1,400 AD & 130,000 Controls) to identify potential genes related to AD.
2. Conducting protein quantitative trait locus (QTL) analysis of 191,440 proteins in serum or plasma to detect associations between AD-associated genotypes and protein abundance.
3. Focusing on the Hispanic population to conduct candidate gene/protein association studies on AD-associated loci suggested by the GWAS and proteomics analysis and identify Hispanic specific AD associated genes/alleles.
4a. Downstream analysis: gene and biological processes including in silico annotation, pathway analysis; fine mapping, integration with eQTL in brain and central nervous system (CNS).
4b. With future/pending funding supports and time, we will validate Hispanic specific AD-associated genes in our own cohorts (N=323) where there are over 96 variables collected.

Anticipated Findings

Overall, this grant proposal seeks to advance our understanding of the genetic basis and protein biomarkers of AD and identify novel therapeutic targets for the disease. By leveraging genomic data and integrating multi-omics approaches, we hope to uncover new insights into AD pathogenesis and ultimately develop more effective treatments for this devastating disorder. We anticipate that this diverse dataset will advance the promise of genomic medicine for all.

Demographic Categories of Interest

Data Set Used

Controlled Tier

Research Team

Owner:

Chun (Cathy) Xu - Mid-career Tenured Researcher, University of Texas, Rio Grande Valley

Collaborators:

Belinda Alvarado - Graduate Trainee, University of Texas, Rio Grande Valley

learning alzheimers

Is there a correlation between Tau protein biomarker testing in patients with Alzheimer’s Disease and healthcare utilization? Across health care access type, are there differences in the correlation between the prevalence of Tao testing and Alzheimer's disease diagnosis Across patients…

Scientific Questions Being Studied

Is there a correlation between Tau protein biomarker testing in patients with Alzheimer’s Disease and healthcare utilization?
Across health care access type, are there differences in the correlation between the prevalence of Tao testing and Alzheimer's disease diagnosis
Across patients with Alzheimer's disease diagnosis, is there a consistent health care utilization pattern

Project Purpose(s)

Educational

Scientific Approaches

secondary data analysis, to enhance learning in PHD curriculum for an assignment and inform future dissertation work

Anticipated Findings

due to the multiple ALz theories, there may not be any correlations between biomarker testing and dx. there may be a relationship with healthcare utilization and diagnosis

Demographic Categories of Interest

This study will not center on underrepresented populations.

Data Set Used

Registered Tier

Research Team

Owner:

carolina frantz - Graduate Trainee, Duquesne University

Pascoal Lab v8 - Douglas

The specific scientific questions we intend to study are: 1. Can we predict the onset of Alzheimer’s Disease and related dementias (ADRD) using machine learning methods applied to electronic health records (EHRs)? 2. What are the major risk factors contributing…

Scientific Questions Being Studied

The specific scientific questions we intend to study are:

1. Can we predict the onset of Alzheimer’s Disease and related dementias (ADRD) using machine learning methods applied to electronic health records (EHRs)?
2. What are the major risk factors contributing to the development of ADRD?
3. Are there identifiable subtypes of ADRD based on cognitive and neuropsychiatric symptoms, and what are their distinct features?
4. How do different subtypes of ADRD progress in terms of disease trajectory, and what is the comparative decline among these subtypes?
5. Can a software tool be developed to predict ADRD and assist clinicians in identifying high-risk patients, improving care and outcomes?

The importance of these questions lies in the potential for early detection and intervention in ADRD, which could significantly alter the disease's impact on patients and the healthcare system.

Project Purpose(s)

Disease Focused Research (Alzheimer's disease)

Scientific Approaches

We will apply machine learning to electronic health records (EHRs) for early prediction of Alzheimer's Disease and related dementias (ADRD), identifying risk factors and subtypes, analyzing disease trajectories, and developing a predictive tool for clinical use.
Datasets: Longitudinal EHR data encompassing demographics, diagnostics, lab results, medications, and clinical notes.
Methods:
• Predictive Modeling: We'll use algorithms like random forests and neural networks for predicting ADRD onset.
• Feature Importance: To pinpoint risk factors, methods such as permutation importance will be applied.
• Clustering: Algorithms like k-means will classify subtypes based on symptoms.
• Trajectory Analysis: Comparative analysis of cognitive decline across subtypes using time-series methods.
Tools:
• Data Processing: Python and R.
• ML Frameworks: scikit-learn, TensorFlow, PyTorch.
• Visualization: Matplotlib, Seaborn, PowerBI.
• Software Development: Python.

Anticipated Findings

The anticipated findings from the study include:
1. A set of predictive markers for ADRD onset derived from EHRs.
2. Identification of major risk factors for ADRD.
3. Distinct subtypes of ADRD based on cognitive and neuropsychiatric symptoms.
4. Differential progression trajectories for each ADRD subtype.
5. A validated software tool that predicts ADRD risk and aids clinical decision-making.
These findings could contribute to the field by:
1. Enhancing early detection capabilities for ADRD, potentially leading to earlier interventions.
2. Providing a deeper understanding of the risk factors, which could inform preventative strategies.
3. Revealing the heterogeneity within ADRD, allowing for more personalized treatment approaches.
4. Offering insights into the progression of ADRD, aiding in the prediction of patient needs.
5. Delivering a practical tool for healthcare professionals, integrating data-driven insights directly into clinical workflows, which may improve patient outcomes.

Demographic Categories of Interest

This study will not center on underrepresented populations.

Data Set Used

Controlled Tier

Research Team

Owner:

Tharick Pascoal - Early Career Tenure-track Researcher, University of Pittsburgh

Alzheimer's Disease and NLR_8

Are blood traits such as neutrophil to lymphocyte ratio associated with Alzheimer's Disease? Neutrophil to lymphocyte ratio has been shown to be associated with Alzheimer's Disease but it is not known whether this association will replicate in the All of…

Scientific Questions Being Studied

Are blood traits such as neutrophil to lymphocyte ratio associated with Alzheimer's Disease? Neutrophil to lymphocyte ratio has been shown to be associated with Alzheimer's Disease but it is not known whether this association will replicate in the All of Us dataset.

Project Purpose(s)

Disease Focused Research (Alzheimer's disease)

Scientific Approaches

The main approach will use a cox proportional hazards model while adjusted for covariates such as age and sex.

Anticipated Findings

Based on the previous literature, it is anticipated that neutrophil to lymphocyte ratio will be associated with Alzheimer's Disease in the All of Us cohort. These findings may allow for a better understanding of Alzheimer's Disease.

Demographic Categories of Interest

This study will not center on underrepresented populations.

Data Set Used

Controlled Tier

Research Team

Owner:

Emily Drzymalla - Other, University of North Carolina, Chapel Hill

Duplicate of Gene-Environment Interactions in Alzheimer’s Disease

60-80% of late-onset Alzheimer’s disease (LOAD) risk is heritable. Both genetic and environmental factors are responsible for the development and progression of LOAD. Many LOAD susceptibility genes have been identified by genome-wide association studies (GWAS). While genetic factors contribute significantly…

Scientific Questions Being Studied

60-80% of late-onset Alzheimer’s disease (LOAD) risk is heritable. Both genetic and environmental factors are responsible for the development and progression of LOAD. Many LOAD susceptibility genes have been identified by genome-wide association studies (GWAS). While genetic factors contribute significantly to the risk of developing LOAD, the impact of environmental factors on the disease remains multifaceted. We propose a research proposal that assesses gene-environment (G×E) interactions in LOAD to analyze All of Us data. We will test these hypotheses: 1) individuals with the same or lower genetic risks will face an increased risk for LOAD when modified by higher environmental risks (vulnerability), whereas those with the same or higher genetic risks will experience a reduced risk when influenced by lower environmental risks (resilience); 2) effects of genetic risks on LOAD will vary modified by midlife and later-life environmental risks across ethnoracial groups.

Project Purpose(s)

Disease Focused Research (Alzheimer's disease)
Population Health
Social / Behavioral
Ancestry

Scientific Approaches

Polygenic risk scores (PRSs) are well-established in LOAD studies and feature an estimate of an individual’s genetic liability to LOAD by aggregating genetic effects of single-nucleotide variants (SNVs), thus serving as a comprehensive score for genetic risks. We will examine the interactions between PRSs and each of environmental risk factors in LOAD. We will also use an item response theory (IRT)-based model to generate environmental risk scores (ERSs) and investigate the interactions between PRSs and midlife/later-life ERSs in LOAD across ethnoracial groups. We will utilize the All of Us genetic data and published GWAS summary statistics to construct LOAD PRSs and construct midlife and later-life ERSs based on various environmental indicators in the All of Us data, including social of determinants of health (e.g., education and incomes), physical conditions (e.g., hypertension, diabetes, and depression), and lifestyle (e.g., smoking, alcohol, and exercise) variables.

Anticipated Findings

Overall, we will examine one-by-one G×E interactions in LOAD, but also construct comprehensive scores, PRSs and midlife/later-life ERSs, for individuals across ethnoracial groups, collectively contributing to advancing our knowledge of G×E interactions on vulnerability and resilience to LOAD. Informed by data indicating an overall risk (PRS modified by ERS) for LOAD, this approach may enable clinicians and individuals to initiate disease screening and discuss life planning strategies. This study would also provide a crucial tool for informing the design and implementation of personalized therapeutic and preventative programs, enabling more precise and individualized approaches to the treatment and prevention of LOAD. This study not only enhances our understanding of progression of LOAD but also provides a foundation for developing targeted strategies to mitigate risk and bolster resilience, thereby advancing personalized interventions for diverse populations at different life stages.

Demographic Categories of Interest

This study will not center on underrepresented populations.

Data Set Used

Controlled Tier

Research Team

Owner:

JIng Zhang - Research Fellow, University of Kentucky

AD disparity

Alzheimer's Disease (AD) is debilitating conditions that impair memory, thought processes, and functioning of millions of primarily older adults, in the US and worldwide. AD arise from complex interactions between genetic (G) and environmental (E) factors with ~70% of risk…

Scientific Questions Being Studied

Alzheimer's Disease (AD) is debilitating conditions that impair memory, thought processes, and functioning of millions of primarily older adults, in the US and worldwide. AD arise from complex interactions between genetic (G) and environmental (E) factors with ~70% of risk of developing AD attributable to genetics and acquired environmental, social and lifestyle risk factors throughout lifespan also contribute significantly to the AD development and progression. Racial disparities exist in AD that older adults from racially minority populations were more likely than white populations to have poorer cognitive functioning, accelerated cognitive decline, and higher AD prevalence. However, the risk factors that contribute to this racial gap and the underlying mechanisms that produce the AD inequality remain largely unknown. In this project, we aim to identify the genetic, social, lifestyle and other environmental risk factors that lead to the racial disparity in AD.

Project Purpose(s)

Disease Focused Research (Alzheimer's disease and psychiatric disorders (e.g. addiction))
Population Health
Methods Development
Ancestry

Scientific Approaches

In this study, we will analyze different types of data (e.g. genetic, multi-omics, survey data, etc.) from a number of racially diverse cohorts including All of Us. For All of Us, we will analyze the genetic data from both white and minority populations and identify the genetic risk factors that drive the racial disparity in AD. We will use trans-ancestry genome-wide association study (GWAS) methods followed by transcriptome-wide association study (TWAS), polygenic risk score analysis and fine mapping tools. We will also analyze the survey questionnaire, electronic health record (EHR) and physical measurement data and apply machine learning methods (e.g. transfer learning methods, neural network based discriminative models such as CNN, GCN, etc.) to identify the social (e.g. SDOH), lifestyle (e.g. LE8) and other environmental risk factors that drive the racial disparity in AD.

Anticipated Findings

In this study, we aim to identify the genetic, social, lifestyle and other environmental risk factors that lead to the racial disparity in AD, with the goal of improving the mechanistic understanding of the aging brain underlying the pathogenesis of AD. In addition, we will also evaluate how health resource utilization, institutionalization and treatment effectiveness of AD differ in different racial groups. The findings will provide us a holistic picture of racial disparity in AD and suggest prevention strategy to lessen the disparity. Our study will also potentially provide possible solutions to clinical implications for reducing AD burden and racial/ethnic AD disparities.

Demographic Categories of Interest

Data Set Used

Registered Tier

Research Team

Owner:

Tianzhou Ma - Mid-career Tenured Researcher, University of Maryland, College Park
menglu liang - Teacher/Instructor/Professor, University of Maryland, College Park

Collaborators:

Gourav Velma - Graduate Trainee, University of Maryland, College Park

Subscribe to Email Updates