Investigating the role of microRNAs in complex traits and disorders: a molecular epidemiological approach
Author(s)
Mustafa, Rima
Type
Thesis
Abstract
MicroRNAs (miRNAs) are small non-coding RNAs that regulate gene expression. Over 2,000 miRNAs have been identified in humans. However, previous studies examining miRNAs in complex disorders were subject to the limited coverage of miRNAs, the number of diseases tested, and the inability to distinguish between correlation and causation. This thesis examines the role of miRNAs in complex traits and disorders using molecular epidemiology approaches to address those limitations.
In the first analysis, phenome-wide association studies were conducted in the UK Biobank (N=423,419) to test the effects of 346 variants in miRNAs on 905 clinical diagnoses. At a false discovery rate (FDR)<0.05, 122 associations for six variants in the seed region of six miRNAs, nine variants in the mature region of six miRNAs, and 24 variants in 27 precursor miRNAs were identified. This analysis highlighted possible clinical importance of variants in miRNAs.
In the second analysis, plasma levels of 2,083 miRNAs were measured in the Rotterdam Study (N=2,178), allowing the conduct of genome-wide association studies to identify miRNA expression quantitative trait loci (miR-eQTLs). At P<2.4×10-11, 3,292 associations between 1,289 miR-eQTLs and 63 miRNAs were discovered. Subsequently, 64.8% of associations were replicated in an independent cohort. Pleiotropic loci, such as in chr9:136128546-136296530, mapped into ABO and nearby genes were identified.
In the third analysis, cis-miR-eQTLs were used to proxy miRNAs and test their associations against clinical diagnoses in the UK Biobank. Single cis-miR-eQTLs were used to proxy 85 miRNAs. Multiple cis-miR-eQTLs were used to compute 119 genetic risk scores (miRNA-GRS). At FDR<0.05, 29 associations for nine cis-miR-eQTLs and 44 associations for 17 miRNA-GRS were identified. Thirty-seven associations were tested using Mendelian randomisation, suggesting potentially causal roles. The associations between miR-329-3p and miR-543 with obesity were replicated. Potential downstream targets were identified, allowing to generate hypotheses for functional studies to dissect underlying mechanisms.
In the first analysis, phenome-wide association studies were conducted in the UK Biobank (N=423,419) to test the effects of 346 variants in miRNAs on 905 clinical diagnoses. At a false discovery rate (FDR)<0.05, 122 associations for six variants in the seed region of six miRNAs, nine variants in the mature region of six miRNAs, and 24 variants in 27 precursor miRNAs were identified. This analysis highlighted possible clinical importance of variants in miRNAs.
In the second analysis, plasma levels of 2,083 miRNAs were measured in the Rotterdam Study (N=2,178), allowing the conduct of genome-wide association studies to identify miRNA expression quantitative trait loci (miR-eQTLs). At P<2.4×10-11, 3,292 associations between 1,289 miR-eQTLs and 63 miRNAs were discovered. Subsequently, 64.8% of associations were replicated in an independent cohort. Pleiotropic loci, such as in chr9:136128546-136296530, mapped into ABO and nearby genes were identified.
In the third analysis, cis-miR-eQTLs were used to proxy miRNAs and test their associations against clinical diagnoses in the UK Biobank. Single cis-miR-eQTLs were used to proxy 85 miRNAs. Multiple cis-miR-eQTLs were used to compute 119 genetic risk scores (miRNA-GRS). At FDR<0.05, 29 associations for nine cis-miR-eQTLs and 44 associations for 17 miRNA-GRS were identified. Thirty-seven associations were tested using Mendelian randomisation, suggesting potentially causal roles. The associations between miR-329-3p and miR-543 with obesity were replicated. Potential downstream targets were identified, allowing to generate hypotheses for functional studies to dissect underlying mechanisms.
Version
Open Access
Date Issued
2022-08-19
Date Awarded
01/12/2022
License URL
Advisor
Dehghan, Abbas
Evangelou, Marina
Sponsor
Imperial College London
Publisher Department
School of Public Health
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)