First consider that (at least for ggplot2) the alpha (opaqueness) parameter is encoded with 8-digit hex color codes, with the last two letters of the eight comprising the alpha channel suffix. This set of strings represents the suffixes that range from 45% thru 100% opaqueness.
alpha_suffix <- c("73", "99","B3","D3","FF")
In your original code, there was a line that I believe needed adjustment. (Because your colors were not lined up with the 'groups'). One 6-digit hex color per group can be had this way:
library(RColorBrewer)
set.seed(1)
id <- rep(1:4, each=5)
group <- rep(LETTERS[1:4], each=5)
inj<- rep(1:5,4)
q1 <- sample(1:50,20)
q2 <- sample(1:50,20)
df <- data.frame(id, group, inj, q1,q2)
group_col <- rep(brewer.pal(4,"Paired"), each=5) # ADJUSTED LINE
Then one can append the alpha suffixes with the 6-digit hex codes, and assign them to the df object as follows. This lines up each color in 'group_col' with the five levels of 'inj'.
df$group_col <- paste0(group_col, rep(alpha_suffix, 4))
By doing it this way, one can use 'group_col' for the color aesthetic, as well as for the values argument in scale_color_manual. Here's what a ggplot2 scatterplot looks like with these colors.
library(ggplot2)
ggplot(df, aes(x= q1, y= q2, color = group_col))+
geom_point(stat="identity", size = 8) +
scale_color_manual(values = df$group_col)+
theme_minimal()
The opacity effects would be easier to see if you try another RColorBrewer palette, like this, instead of the "Paired":
group_col <- rep(brewer.pal(4,"Set1"), each=5)
And finally, in base R 'group_col' could be implemented like this:
plot(df$q1,df$q2,col= df$group_col,pch=16,cex=4)
