Cell-type-specific analysis in Alzheimer’s disease Open Access

Abstract

Background: 

Alzheimer’s disease (AD) is one of the chronic neurodegenerative diseases which starts slowly but worsens over time. It has been found that distinct cell types in the brain play different roles in AD etiology and progression. 

 

Methods and Materials: 

We estimated cell-type-specific proportions in the brain applying reference-free DNA methylation mixture deconvolution (R/CRAN package RefFreeEWAS) and the newly developed TOAST (Tools for the Analysis of Heterogeneous Tissues, an R/Bioconductor package) statistical method. We detected cell-type specific differential signals between patients with Alzheimer’s and normal controls. Differential methylation calling without consideration of cell mixtures was conducted to compare the results using Bioconductor package Minfi. We conducted the downstream analysis, including pathway/GO analysis and motif analysis using EnrichR

 

Results: 

We assumed there are four brain cell types which are neuron, astrocyte, oligo, microglia during our analysis. We mainly focus on cell type 3 since both Reference-free DNA methylation mixture deconvolution and TOAST statistical method express the lowest average proportion on cell type 3. The average proportion of cell type 3 is 0.0809 (SD=0.0049) for AD patients and 0.0803 (SD=0.0042) for control groups resulted from the Reference-free DNA methylation mixture deconvolution. And the average proportion of cell type 3 generated from the TOAST statistical method is 0.0882 (SD=0.0043) for AD patients and 0.0803 (SD=0.0035) for control groups. The proportions of cell type 1 and 2 are highly negatively correlated to each other with correlation coefficients -0.94 indicating the possible cell type 1 and 2 might be neuron and oligo which are anti-correlated conducted from a recently published paper. The number of DMCs of cell type 3 identified by TOAST is less than that of RefFreeEWAS although it has the least proportions. The biological process regulation of adiponectin secretion has the highest combined score and the neurotransmitter receptor complex is the top cellular component GO term shown in the GO analysis. 

 

Conclusion: 

The study reveals that differential analysis could provide insights for diagnostic biomarkers and therapeutic targets and both RefFreeEWAS and TOAST approaches should be encouraged in the practice that future studies could address cell-type-specific idea in AD area.

Table of Contents

Introduction    1

Methods         4

Description of subject and data         4

Proportions Estimation           5

Cell-type specific DM (csDM) test using TOAST         7

Pathway and Gene Ontologies Analysis         9

Results 10

Data Description         10

Estimated cell Proportions     10

DMCs of four unknown cell types in brain, and joint signal analysis 15

Comparison with existing results       16

Pathway Analysis and GO Ontologies Analysis          18

Discussion       20

Conclusion      21

Reference       22

Appendix         24

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	Statistics
Keyword	Alzheimer’s disease Gene expression DNA methylation Cell-type composition Reference-free deconvolution
Committee Chair / Thesis Advisor	Wu, Hao, Emory University

Last modified

Primary PDF

Thumbnail	Title	Date Uploaded	Actions
	Cell-type-specific analysis in Alzheimer’s disease ()	2020-04-15 14:47:47 -0400	Press to Select an action Download

Supplemental Files

Thumbnail	Title	Date Uploaded	Actions