Accounting for multiple comparisons in a genome-wide association study (GWAS)

Abstract

BackgroundAs we enter an era when testing millions of SNPs in a single gene association study will become the standard, consideration of multiple comparisons is an essential part of determining statistical significance. Bonferroni adjustments can be made but are conservative due to the preponderance of linkage disequilibrium (LD) between genetic markers, and permutation testing is not always a viable option. Three major classes of corrections have been proposed to correct the dependent nature of genetic data in Bonferroni adjustments: permutation testing and related alternatives, principal components analysis (PCA), and analysis of blocks of LD across the genome. We consider seven implementations of these commonly used methods using data from 1514 European American participants genotyped for 700,078 SNPs in a GWAS for AIDS.ResultsA Bonferroni correction using the number of LD blocks found by the three algorithms implemented by Haploview resulted in an insufficiently conservative threshold, corresponding to a genome-wide significance level of α = 0.15 - 0.20. We observed a moderate increase in power when using PRESTO, SLIDE, and simpleℳ when compared with traditional Bonferroni methods for population data genotyped on the Affymetrix 6.0 platform in European Americans (α = 0.05 thresholds between 1 × 10-7 and 7 × 10-8).ConclusionsCorrecting for the number of LD blocks resulted in an anti-conservative Bonferroni adjustment. SLIDE and simpleℳ are particularly useful when using a statistical test not handled in optimized permutation testing packages, and genome-wide corrected p-values using SLIDE, are much easier to interpret for consumers of GWAS studies.

Genome-wide association studies of doubled haploid exotic introgression lines for root system architecture traits in maize (Zea mays L.)

“…Multiple testing in GWAS was accounted for using the statistical program simpleM [29,30], implemented in R [28], which calculates the number of informative SNPs (M eff_G ). First, a correlation matrix for all markers was constructed, and the corresponding eigenvalues for each SNP locus were calculated.…”

“…Plant Science 268 (2018) [30][31][32][33][34][35][36][37][38] conditions. Nonetheless, Abdel-Ghani et al [2] found significant and positive correlations between seedling root and adult plant traits, indicating that more vigorous seedling growth might contribute to a higher grain yield.…”

Simple Statistical Tools to Detect Signals of Recent Polygenic Selection

“…Frequencies of 89 SNPs from the Giant Consortium 7 with a P value = < 1 × 10^-8 (corresponding to P = 0.01 after correction for multiple comparisons, see Johnson et al, 2010) were obtained from 1,000 Genomes. A one-way ANOVA was performed to test for different frequencies of height increasing alleles across the three human races (African, East Asian, European).…”

Part 1: Background, Methodology, and Clinical Adoption of Pharmacogenetics

“…Because up to hundreds of thousands of SNPs are analyzed, genomewide association study (GWAS) requires the use of advanced bioinformatics to handle the extensive amount of data. In addition, because of multiple comparisons within a GWAS, the data need to be corrected for multiple testing [42]. When hundreds of thousands of comparisons are performed, the p-value to reach significance may become too stringent; this is a common phenomenon in GWAS and is an important cause for replication failure.…”