TreeQTL: hierarchical error control for eQTL findings

Peterson, Christine B.; Bogomolov, Marina; Benjamini, Yoav; Sabatti, Chiara

doi:10.1093/bioinformatics/btw198

Cited by 37 publications

(38 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We added sex, genotype principal components (maximum of three), and surrogate variables sequentially. After multiple hypothesis correction, the number of eAssociations (defined as a SNP-gene pair that passed hierarchical multiple hypothesis testing by TreeQTL 46 ) increased monotonically as the number of covariates, consistent with our intuition that sva only returns significant and independent surrogate variables. Therefore, we decided to use sex, the top three genotype principal components and all surrogate variables (four and five for glucose and galactose conditions, respectively).…”

Section: Expression Qtl Mapping and Quality Controlsupporting

confidence: 83%

“…To determine the genetic effects on gene expression in fRPE, we used RASQUAL 45 to map eQTL by leveraging both gene-level and allele-specific count information to boost discovery power. Multiple-hypothesis testing for both glucose and galactose conditions was conducted jointly with a hierarchical procedure called TreeQTL 46 . At FDR < 0.05, we found 726 shared, 272 glucose-specific, and 191 galactose-specific eQTL ( Table 1, S10 and S11, Fig.…”

Section: Expression and Splicing Quantitative Trait Loci Discoverymentioning

confidence: 99%

See 1 more Smart Citation

Ocular disease mechanisms elucidated by genetics of human fetal retinal pigment epithelium gene expression

Liu

Calton

Benchorin

et al. 2018

Preprint

View full text Add to dashboard Cite

The eye is an intricate organ with limited representation in large-scale functional genomics datasets. The retinal pigment epithelium (RPE) serves vital roles in ocular development and retinal homeostasis. We interrogated the genetics of gene expression of cultured human fetal RPE (fRPE) cells under two metabolic conditions. Genes with disproportionately high fRPE expression are enriched for genes related to inherited ocular diseases. Variants near these fRPE-selective genes explain a larger fraction of risk for both age-related macular degeneration (AMD) and myopia than variants near genes enriched in 53 other human tissues.Increased mitochondrial oxidation of glutamine by fRPE promoted expression of lipid synthesis genes implicated in AMD. Expression and splice quantitative trait loci (e/sQTL) analysis revealed shared and metabolic condition-specific loci of each type and several eQTL not previously described in any tissue. Fine mapping of fRPE e/sQTL across AMD and myopia genome-wide association data suggests new candidate genes, and mechanisms by which the same common variant of RDH5 contributes to both increased AMD risk and decreased myopia risk. Our study highlights the unique transcriptomic characteristics of fRPE and provides a resource to connect e/sQTL in a critical ocular cell type to monogenic and complex eye disorders.3The importance of vision to humans and the accessibility of the eye to examination have motivated the characterization of more than one thousand genetic conditions involving ocular phenotypes 1 . Among these, numerous monogenic diseases exhibit significant inter-and intra-familial phenotypic variability 2-7 . Imbalance in allelic expression of a handful of causative genes has been documented 8 , but common genetic variants responsible for such effects remain to be discovered.Complementing our knowledge of numerous monogenic ocular disorders, recent genome-wide association studies (GWAS) 9 have identified hundreds of independent loci associated with polygenic ocular phenotypes such as age-related macular degeneration (AMD), the leading cause of blindness in elderly individuals in developed countries 10,11 , and myopia, the most common type of refractive error worldwide and an increasingly common cause of blindness 12-14 . Despite the rapid success of GWAS in mapping novel ocular disease susceptibility loci, the functional mechanisms underlying these associations are often obscure.Connecting changes in molecular functions such as gene expression and splicing with specific GWAS genomic variants has aided the elucidation of functional mechanisms. Non-coding variants account for a preponderance of the most significant GWAS loci 15,16 , and most expression quantitative trait loci (eQTL) map to non-coding variants 17 . Thousands of eQTL have been found in a variety of human tissues 18 , but ocular celltypes are underrepresented among eQTL maps across diverse tissues, and no systematic search for eQTL has so far been described for any cell type in the human eye.The retinal pigment epithelium ...

show abstract

Section: Expression Qtl Mapping and Quality Controlsupporting

confidence: 83%

Section: Expression and Splicing Quantitative Trait Loci Discoverymentioning

confidence: 99%

Ocular disease mechanisms elucidated by genetics of human fetal retinal pigment epithelium gene expression

Liu

Calton

Benchorin

et al. 2018

Preprint

View full text Add to dashboard Cite

show abstract

“…Third, cis-pQTLs in linkage equilibrium that did not meet the trans-pQTL significance threshold in Sun et al were identified in a step-forward procedure. To identify additional cis-pQTLs we applied a hierarchical multiple testing correction procedure that has been shown to control the false discovery rate at 5% in expression quantitative trait loci (eQTL) studies 72,73 to the pQTL summary statistics from Sun et al 2018 15 . For each aptamer, P-values within each 1MB of any gene encoding the targeted protein or protein complex were first Bonferroni corrected for the number of tests within that cis window(s) to obtain locally corrected P-values.…”

Section: Mendelian Randomisation Analysesmentioning

confidence: 99%

Integrative analysis of the plasma proteome and polygenic risk of cardiometabolic diseases

Ritchie

Liu

Teo

et al. 2019

Preprint

View full text Add to dashboard Cite

Polygenic risk scores (PRSs) capture the polygenic architecture of common diseases by aggregating genome-wide genetic variation into a single score that reflects individual's disease risk, affording a new opportunity to identify downstream molecular pathways involved in disease pathogenesis. We performed an integrative analysis to characterise associations of PRSs of five cardiometabolic diseases with 3,442 plasma proteins in 3,175 healthy individuals. Through polygenic association scans we identified 48 plasma proteins whose levels were associated with PRSs for coronary artery disease (CAD), chronic kidney disease (CKD), or type 2 diabetes (T2D). This approach identified both well-known disease-associated proteins as well as those with previously no known link to these diseases. We found that PRSs to protein associations were largely truly polygenic; independent of single loci and genomic regions expected to have strong effects on protein levels. Our integrative analysis and laboratory experiments revealed a role for polygenic effects on several well-known disease proteins and identified several promising novel targets for follow-up studies, including genes which through Mendelian randomization analysis displayed causal evidence for effects on disease risk. Mouse studies highlighted specific tissues and phenotypes for PRSassociated human proteins. We found that implicated genes were responsive to dietary intervention in mice and showed strong evidence of druggability in humans, consistent with PRS-associated proteins having therapeutic potential. Overall, our study provides a framework for polygenic association studies, and demonstrates the power of polygenic scores to unravel novel disease biology. Research (NIHR), the NIHR BioResource

show abstract

“…The hierarchical error control methods described in this paper have been implemented as a part of the TreeQTL R package [Peterson, ], available online at http://bioinformatics.org/treeqtl.…”

Section: Softwarementioning

confidence: 99%

Many Phenotypes Without Many False Discoveries: Error Controlling Strategies for Multitrait Association Studies

Peterson

Bogomolov

Benjamini

et al. 2015

Genetic Epidemiology

View full text Add to dashboard Cite

The genetic basis of multiple phenotypes such as gene expression, metabolite levels, or imaging features is often investigated by testing a large collection of hypotheses, probing the existence of association between each of the traits and hundreds of thousands of genotyped variants. Appropriate multiplicity adjustment is crucial to guarantee replicability of findings, and the false discovery rate (FDR) is frequently adopted as a measure of global error. In the interest of interpretability, results are often summarized so that reporting focuses on variants discovered to be associated to some phenotypes. We show that applying FDR‐controlling procedures on the entire collection of hypotheses fails to control the rate of false discovery of associated variants as well as the expected value of the average proportion of false discovery of phenotypes influenced by such variants. We propose a simple hierarchical testing procedure that allows control of both these error rates and provides a more reliable basis for the identification of variants with functional effects. We demonstrate the utility of this approach through simulation studies comparing various error rates and measures of power for genetic association studies of multiple traits. Finally, we apply the proposed method to identify genetic variants that impact flowering phenotypes in Arabidopsis thaliana, expanding the set of discoveries.

show abstract

TreeQTL: hierarchical error control for eQTL findings

Cited by 37 publications

References 13 publications

Ocular disease mechanisms elucidated by genetics of human fetal retinal pigment epithelium gene expression

Ocular disease mechanisms elucidated by genetics of human fetal retinal pigment epithelium gene expression

Integrative analysis of the plasma proteome and polygenic risk of cardiometabolic diseases

Many Phenotypes Without Many False Discoveries: Error Controlling Strategies for Multitrait Association Studies

Contact Info

Product

Resources

About