How to add a pseudo aesthetic argument to ggplot2 geom?

Question

What I have and what I want:

data<- data.frame(value= 1:10, Treatment= rep(c("A", "B"),5), Species=rep(c("alpha", "beta"), each=5))

And it is known to all that ggplot2::aes() function can automatically group the data according to the arguments it quotes, for example, ggplot(data=data)+geom_boxplot(aes(y= value, x= Treatment, fill=Treatment)) separates and groups data by Treatment

and is plotted as 2 different boxes. In like manner ggplot(data=data)+geom_boxplot(aes(y= value, x= Treatment, fill=Treatment, linetype= Species)) will generate 4 boxes grouped by both Teatment and Species with different color filling and linetype respectively .

However, how to generate a boxplot of 4 boxes grouped by Treatment and Species in which Treatment is presented by filling color same as the plot metioned above, but Species was not presented, that is, the linetype is all the same among boxes? In other words, is there a null aesthetic only group the data but never be presented on the plot.

Answer 1

You mean this?

ggplot(data=data)+
  geom_boxplot(aes(y= value, x= Treatment, fill=Treatment)) + 
  facet_wrap(~Species)

Answer 2

You can define a new variable that has 4 levels either in your data or on the fly as I have shown below. With this you will get the 4 box plots, colored and grouped as you asked for. No matter what you will probably need to distinguish the data in some way, either on the axis labels or in the legend.

ggplot(data=data)+geom_boxplot(aes(y= value, x= Treatment, fill=Treatment, group=paste(Species, Treatment))

Answer 3

结合J Hart和baptiste的解决方案， ggplot(data=data)+geom_boxplot(aes(y= value, x= Treatment, fill=Treatment, group= paste(Treatment, Species)))完美工作。