Boxplots aren't colouring or plotting labels properly in R, why? - r

Question

My Tukey test significant results LABELS and the colours plotted as box plots do not plot over each sample box plot. Why?
Seems like the labels are plotted at different y-axis along the same s1 (x-axis)?
Reproducible dataset here:
library(multcompView)
df <- data.frame('Sample'=c("s1","s1","s1","s1","s1","s2","s2","s2","s2","s2","s3","s3","s3","s3","s4","s4","s5","s5"), 'value'=c(-0.1098,-0.1435,-0.1046,-0.1308,-0.1523,-0.1219,-0.1114,-0.1328,-0.1589,-0.1567,-0.1395,-0.1181,-0.1448,-0.124,-0.1929,-0.1996,-0.1981,-0.1917))
anova_df <- aov(df$value ~ df$Sample )
tukey_df <- TukeyHSD(anova_df, 'df$Sample', conf.level=0.95)
# I need to group the treatments that are not different each other together.
TUKEY <- tukey_df
generate_label_df <- function(TUKEY, variable){
# Extract labels and factor levels from Tukey post-hoc
Tukey.levels <- TUKEY[[variable]][,4]
Tukey.labels <- data.frame(multcompLetters(Tukey.levels)['Letters'])
#I need to put the labels in the same order as in the boxplot :
Tukey.labels$Sample=rownames(Tukey.labels)
Tukey.labels=Tukey.labels[order(Tukey.labels$Sample) , ]
return(Tukey.labels)
}
# Apply the function on my dataset
LABELS <- generate_label_df(TUKEY , "df$Sample")
# A panel of colors to draw each group with the same color :
my_colors <- c(
rgb(143,199,74,maxColorValue = 255),
rgb(242,104,34,maxColorValue = 255),
rgb(111,145,202,maxColorValue = 255))
# Draw the basic boxplot
a <- boxplot(df$value ~ df$Sample , ylim=c(min(df$value) , 1.1*max(df$value)) , col=my_colors[as.numeric(LABELS[,1])] , ylab="Value" , main="")
# I want to write the letter over each box. Over is how high I want to write it.
over <- 0.1*max(a$stats[nrow(a$stats),] )
#Add the labels
text(c(1:nlevels(df$Sample)), a$stats[nrow(a$stats),]+over, LABELS[,1] , col=my_colors[as.numeric(LABELS[,1])] )
Current output:
Desired plot-like (colours and LABELS):

Answer 1

First, LABELS$Letters is a character vector. You can get as.numeric(LABELS[,1]) to work if you make it a factor first.
Second, your y-limit needs some work for negative values. There is a function you might find useful called extendrange which is used in many a plotting function.
This line c(1:nlevels(df$Sample)) also would work if df$Sample was a factor which is was not.
Also, if you are plotting text at a specific location, you can adjust the text using either text(..., pos = ) or text(..., adj = ) to shift the position.
LABELS$Letters <- factor(LABELS$Letters)
a <- boxplot(df$value ~ df$Sample , ylim = extendrange(df$value), col=my_colors[as.numeric(LABELS[,1])] , ylab="Value" , main="")
text(seq_along(a$names), apply(a$stats, 2, max), LABELS[,1], col=my_colors[as.numeric(LABELS[,1])], pos = 3)

Answer 2

If you don't mind changing your workflow and use tidyverse library this is how you could achieve your goal:
# join df and LABELS into one data table
inner_join(df, LABELS, by = "Sample") %>%
# calculate max value for each Sample group (it will be used to place the labels)
group_by(Sample) %>%
mutate(placement = max(value)) %>%
ungroup() %>%
# make a plot
ggplot(aes(Sample, value, fill = Letters))+
geom_boxplot()+
geom_text(aes(y = placement, label = Letters, col = Letters), nudge_y = 0.01, size = 6)+
theme_minimal()+
theme(legend.position = "none")