R-根据p值的颜色图

Question

I have hundreds of boxplots that are printed automatically from my huge dataset and I have fabricated the code to colour the plots 'coral' if the variable has a statistical significance across the selected groups and 'aquamarine' if significance is not detected. This I did with col=ifelse(...< 0.05,'coral','aquamarine') .

I though my output would be much easier to skim through if the boxplots were coloured according to the p-value. So I would like to tell col=... to colour plots from variable with p-value in the range 0.05 - 0.01 to yellow, 0.01 - 0.005 to orange and <0.005 to red.

I tried this with

boxplot(a~b, data=df, col=if(pv1<0.05)"yellow" else if (pv1<0.01) "orange" else if (pv1<0.005) "red" else "green")

As you can see that this wont do because the values overlap. What I need to do is write the ranges inside the if sentence but I just dont know how to.

Here is some data to work with. Note that the actual p-values of the groups are not the same as give here. I just display pv1, 2 and 3 so you can test out the code.

a <- c(23,24,64,12,4,75,12,65,86,76)
b <- c(1,2,2,1,2,1,3,3,1,3)
df  <-  data.frame(a,b)
pv1  <- c(0.05)
pv2  <- c(0.01)
pv3  <- c(0.005)

Answer 1

To modify your if...else statement, you should think it the other way around (here, you're asking that "if it is not < 0.05, then if it is < 0.01..." which is not possible), so you can write:

col=if(pv1<0.005)"red" else if (pv1<0.01) "orange" else if (pv1<0.05) "yellow" else "green"

Example

pv1 <- 0.04 ; if(pv1<0.005)"red" else if (pv1<0.01) "orange" else if (pv1<0.05) "yellow" else "green" # "yellow"
pv1 <- 0.004 ; if(pv1<0.005)"red" else if (pv1<0.01) "orange" else if (pv1<0.05) "yellow" else "green" # "red"
pv1 <- 0.06 ; if(pv1<0.005) "red" else if (pv1<0.01) "orange" else if (pv1<0.05) "yellow" else "green" # "green"
pv1 <- 0.006 ; if(pv1<0.005) "red" else if (pv1<0.01) "orange" else if (pv1<0.05) "yellow" else "green" # "orange"

Another option , if your pvalues are in a vector, you can try with cut :

mycolours <- as.character(cut(pv1, ,c(1,0.05,0.01,0.005,0), right=F, labels=c("red","orange","yellow","aquamarine"), include.lowest=T))

And then

boxplot(a~b, data=df, col=mycolours)