The meaning of q-q plot in Gwas _plot

Q-q plot is quantile-quantile plot. It is often used in various types of research, mainly to visually indicate the difference between observed and predicted values.

In the SPSS is very tolerant to do, analysis-descriptive Statistics-q-qplot.

Q-q plot is mainly used to estimate the difference between the observed and predicted values of quantitative traits. In general, the quantitative trait data we obtained are normal distribution data. In the GWAS study, the x and y axes of q-q plot are mainly representative of-LG P values of each SNP. The predicted line is a line of 45° at the point of origin. The actual observation value is the marked solid point.

Q-q Plot main points:

Why is the predicted dashed line 45° out of it? Because the line of prediction is actually drawn by the first quadrant in the QQ chart. Theoretically a point A on the map position should be a predictive value =a the actual value, the conversion to coordinates is a (x,y) x=y. So the predicted line is a 45° line that is emitted from the origin.

The coordinates of the points of the observed values are not. The coordinates of the same position A are (x,y) x is the predicted value and Y is the actual observed value. Check the R in QQ plot algorithm is like this

Pvals <-read.table ("Dgi_chr3_pvals.txt", header=t)

observed <-sort (pvals$pval)
LOBs <--( LOG10 (observed))

expected <-C (1:length (observed))
Lexp <--(log10 (expected) (length (expected) +1))

The specific explanation is that the P value is sorted first from small to large. The lobs represents the ordinate, the lexp represents the horizontal axis, the ordinate is the-log10 that observes the P value, and the horizontal axis depends on the number of P values. For example, when there are only 3 p values p1=0.0001 p2=0.001 p3=0.01, then in this P-value group, length (observed) = 3, for p1=0.0001 expected=1 lexp=-log10 (1/3+1), for p2= 0.001 expected=2 lexp=-log10 (2/3+1), p3=0.01 expected=3 (LEXP=-LOG10) ..... So If there is a deviation indicating that the actual value is biased against the predicted value, and that the SNP point has a large deviation in the GWAS study, the deviation from the observed value of this SNP is caused by the genetic effect of the SNP mutation.