Cell-type-specific Transcriptomic-wide Association Studies Detected 80 Independent Risk Genes for Alzheimer’s Disease Dementia Restricted; Files Only

Abstract

Most Transcriptome-wide Association Studies (TWASs) of Alzheimer’s disease (AD) dementia utilize only bulk-level RNA-seq data of dorsolateral prefrontal cortex (DLPFC) with GWAS summary data, which often misses risk genes with cell-type-specific effects. Here, we utilize our recently proposed omnibus xWAS (xWAS-O) pipeline to leverage a large scale study of snRNA-seq data of DLPFC (n=436) to detect cell-type-specific risk genes of AD dementia. Applying xWAS-O respectively to six major brain cell types, we identified 235 cell-type-specific TWAS risk genes, including 80 fine-mapped independent risk genes of which 9 are novel findings. By the analogous PWAS-O pipeline using proteomics data of DLPFC of the same study cohort (n=716) and the same GWAS summary data, we identified 35 significant PWAS risk genes of AD dementia, including 19 fine-mapped independent significant PWAS risk. Additionally, we found 32 out of 80 independent cell-type-specific TWAS-O risk genes overlap with PWAS risk genes, and 13 of them were further validated with PWAS results using Bonferroni corrected threshold. Further, protein-protein interaction (PPI) network analyses revealed functional enrichment in phenotypes like family history of Alzheimer’s disease and Alzheimer’s disease biomarker measurements, underscoring the role of different cell types in AD pathogenesis. Notably, we identified 2 validated independent risk genes and 1 enriched AD-related pathway specific to microglia. Our findings highlight the importance of integrating the novel large-scale snRNA-seq and proteomics data of DLPFC with GWAS summary data to uncover novel molecular mechanisms underlying AD dementia.

Table of Contents

Chapter 1. Introduction  1

Chapter 2. Methods  5

2.1. xWAS-O analytical framework  5

2.2. Fine-map xWAS-O risk genes by GIFT  7

2.3. PPI network and enrichment analyses by STRING  9

2.4. GWAS summary data of AD dementia by Bellenguez et. al. 2022  9

Chapter 3. Results  10

3.1. Methods Overview  10

3.1.1. xWAS-O pipeline  10

3.1.2. ROS/MAP omics data  11

3.1.3. Apply xWAS-O pipeline to ROS/MAP data  13

3.2. Cell-type-specific TWAS-O Results  14

3.3. Comparison of PWAS-O and Cell-type-specific TWAS-O Findings  17

3.3.1. PWAS-O results  17

3.3.2. Compare to cell-type-specific TWAS-O findings  18

3.3.3. PPI Network Analyses  20

Chapter 4. Discussion  26

Appendix A. Figures  31

Figure 1. Overview of xWAS-O pipeline  31

Figure 2. Manhattan plots (-log10(TWAS-O p-values)) by xWAS-O of AD dementia in astrocytes, excitatory neurons, inhibitory neurons, microglia, oligodendrocytes and oligodendrocyte precursor cells  32

Figure 3. PWAS-O results of AD dementia  33

Figure 4. Fine mapping results of TWAS results of AD dementia by GIFT in Microglia  34

Figure 5. PPI network and enrichment analyses results with TWAS risk genes of AD dementia in Astrocytes, Excitatory Neurons, Inhibitory Neurons, Microglia, Oligodendrocytes and Oligodendrocyte Precursor Cells by STRING  35

Appendix B. Tables  36

Table 1. The total number of tested genes/proteins for studying AD dementia by three individual tools  36

Table 2. Independent significant TWAS risk genes identified by xWAS-O in astrocytes (Ast) and microglia (Mic)  38

Table 3. Independent significant TWAS risk genes identified by xWAS-O in excitatory neurons (Ex) and inhibitory neurons (In)  40

Table 4. Independent significant TWAS risk genes identified by xWAS-O in oligodendrocytes (Oli) and oligodendrocyte precursor cells (Opc)  42

Appendix C. Supplementary Figures  44

Fig. S1. Pairwise comparison of CV of cell-type-specific gene expression imputation models by TIGAR/DPR, PrediXcan/EN, and FUSION/BestModel in astrocytes and microglia.  44

Fig. S2. Pairwise comparison of CV of cell-type-specific gene expression imputation models by TIGAR/DPR, PrediXcan/EN, and FUSION/BestModel in Excitatory Neurons and Inhibitory Neurons.  45

Fig. S3. Pairwise comparison of CV of cell-type-specific gene expression imputation models by TIGAR/DPR, PrediXcan/EN, and FUSION/BestModel in Oligodendrocytes and Oligodendrocyte Precursor Cells.  46

Fig. S4. Pairwise comparison of CV of protein abundance imputation models by TIGAR/DPR, PrediXcan/EN, and FUSION/BestModel in Proteomics.  47

Fig. S5. Quantile-Quantile (Q-Q) plots of transcriptome-wide -log10(p-values) obtained by TWAS-O and three individual tools for studying AD dementia in astrocytes.  48

Fig. S6. Quantile-Quantile (Q-Q) plots of transcriptome-wide -log10(p-values) obtained by TWAS-O and three individual tools for studying AD dementia in microglia.  49

Fig. S7. Quantile-Quantile (Q-Q) plots of transcriptome-wide -log10(p-values) obtained by TWAS-O and three individual tools for studying AD dementia in excitatory neurons.  50

Fig. S8. Quantile-Quantile (Q-Q) plots of transcriptome-wide -log10(p-values) obtained by TWAS-O and three individual tools for studying AD dementia in inhibitory neurons.  51

Fig. S9. Quantile-Quantile (Q-Q) plots of transcriptome-wide -log10(p-values) obtained by TWAS-O and three individual tools for studying AD dementia in oligodendrocytes.  52

Fig. S10. Quantile-Quantile (Q-Q) plots of transcriptome-wide -log10(p-values) obtained by TWAS-O and three individual tools for studying AD dementia in oligodendrocyte precursor cells.  53

Fig. S11. Quantile-Quantile (Q-Q) plots of transcriptome-wide -log10(p-values) obtained by PWAS-O and three individual tools for studying AD dementia in proteomics.  54

Fig. S12. Locus plots of the fine-mapping results by GIFT of TWAS-O results in astrocytes (Ast).  55

Fig. S13. Locus plots of the fine-mapping results by GIFT of TWAS-O results in microglia (Mic).  56

Fig. S14. Locus plots of the fine-mapping results by GIFT of TWAS-O results in excitatory neurons (Ex).  57

Fig. S15. Locus plots of the fine-mapping results by GIFT of TWAS-O results in inhibitory neurons (In).  58

Fig. S16. Locus plots of the fine-mapping results by GIFT of TWAS-O results in oligodendrocytes (Oli).  59

Fig. S17. Locus plots of the fine-mapping results by GIFT of TWAS-O results in oligodendrocyte precursor cells (Opc).  60

Fig. S18. Locus plots of the fine-mapping results by GIFT of TWAS-O results in proteomics (Proteome).  61

Fig. S19. Jaccard similarity indexes of overlapped transcriptome-wide/proteome-wide significant risk genes detected by pair-wise analyses.  62

Appendix D. Supplementary Tables  63

Table S1. TWAS-O results in Bulk  63

Table S2. TWAS-O results in Proteome  66

Table S3. TWAS-O results in Astrocytes  68

Table S4. TWAS-O results in Microglia  71

Table S5. TWAS-O results in Excitatory Neurons  74

Table S6. TWAS-O results in Inhibitory Neurons  77

Table S7. TWAS-O results in Oligodendrocytes  80

Table S8. TWAS-O results in Oligodendrocyte Precursor Cells  83

References  86

About this Master's Thesis

Rights statement

• Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.

School	Rollins School of Public Health
Department	Biostatistics
Subfield / Discipline	Biostatistics - MPH & MSPH
Degree	M.S.P.H.
Submission	Master's Thesis
Language	English
Research Field	Biology, Genetics Biology, Biostatistics Statistics
Palavra-chave	Biostatistics Genomics
Committee Chair / Thesis Advisor	Jingjing Yang, Emory University
Committee Members	Michael Epstein, Emory University

Última modificação

Primary PDF

Thumbnail	Title	Date Uploaded	Actions
	File download under embargo until 15 May 2027	2025-04-14 19:32:49 -0400	File download under embargo until 15 May 2027

Supplemental Files

Thumbnail	Title	Date Uploaded	Actions