major allele as the “effect SNP.” Thus, if you have a negative β, the
major allele of that SNP position has a negative effect on your
phenotype, whereas the minor allele could have the opposite effect.
The p-value indicates the significance of the association, and the pvalue threshold will be corrected for the multiple hypothesis test as
described in the following sections. SNPs with “-nan” should be
ignored since their significance cannot be calculated.
3.6 Evaluation of
GWAS Results with
Quantile–Quantile Plot
(Q–Q Plot).
A Q–Q plot allows you to examine whether there were any confounding factors causing false associations between SNPs and the
phenotype. The x-axis of the Q–Q plot denotes the expected
association significance of the test statistics, and the y-axis denotes
the association significance of actual SNPs. If there are no significant confounding effects, the Q–Q plot will not deviate from the
X¼Y axis and only SNPs with strong association with the phenotype will deviate, with a sharp increase at the end of the curve
indicating a true association. The R package “qqman” has functions
to plot Q–Q plots for the given GWAS results. The .ps result file
from EMMAX can be used as input data.
3.6.1 Run R and Install
qqman with the Following
Command
>install.packages("qqman")
3.6.2 Run R and the
Plotting Process
1. Load qqman library.
>library(qqman)
2. Read the input data file.
>ps<-na.omit(read.table('Ath_glucose_germrate_emmax.ps',sep='\t'))
>colnames(ps)<-c('SNP','beta','pvalue')
3. Plot the input data.
>png(file='glucose_emmax_qqplot.png',width=12,height=12,units='in',res=500)#out file
in png format
>par(mar=c(6.1,6.1,2.1,2.1)) #give margins
>qq(ps$pvalue) #plot qqplot
>dev.off()
The output is a “glucose_emmax_qqplot.png” image file in a .
png file format. This file can also be saved in other image formats, as
for the PCA plot described above. The resulting plot (Fig. 2) shows
that the level of significance value deviates from the X¼Y axis. If the
Genome-Wide Association Studies in Arabidopsis
197
major allele of that SNP position has a negative effect on your
phenotype, whereas the minor allele could have the opposite effect.
The p-value indicates the significance of the association, and the pvalue threshold will be corrected for the multiple hypothesis test as
described in the following sections. SNPs with “-nan” should be
ignored since their significance cannot be calculated.
3.6 Evaluation of
GWAS Results with
Quantile–Quantile Plot
(Q–Q Plot).
A Q–Q plot allows you to examine whether there were any confounding factors causing false associations between SNPs and the
phenotype. The x-axis of the Q–Q plot denotes the expected
association significance of the test statistics, and the y-axis denotes
the association significance of actual SNPs. If there are no significant confounding effects, the Q–Q plot will not deviate from the
X¼Y axis and only SNPs with strong association with the phenotype will deviate, with a sharp increase at the end of the curve
indicating a true association. The R package “qqman” has functions
to plot Q–Q plots for the given GWAS results. The .ps result file
from EMMAX can be used as input data.
3.6.1 Run R and Install
qqman with the Following
Command
>install.packages("qqman")
3.6.2 Run R and the
Plotting Process
1. Load qqman library.
>library(qqman)
2. Read the input data file.
>ps<-na.omit(read.table('Ath_glucose_germrate_emmax.ps',sep='\t'))
>colnames(ps)<-c('SNP','beta','pvalue')
3. Plot the input data.
>png(file='glucose_emmax_qqplot.png',width=12,height=12,units='in',res=500)#out file
in png format
>par(mar=c(6.1,6.1,2.1,2.1)) #give margins
>qq(ps$pvalue) #plot qqplot
>dev.off()
The output is a “glucose_emmax_qqplot.png” image file in a .
png file format. This file can also be saved in other image formats, as
for the PCA plot described above. The resulting plot (Fig. 2) shows
that the level of significance value deviates from the X¼Y axis. If the
Genome-Wide Association Studies in Arabidopsis
197