Different font faces within a single legend

Karolina Maciag

unread,

Jul 8, 2015, 4:08:32 PM7/8/15

to ggp...@googlegroups.com

Is it possible to italicize only certain legend entries in a ggplot?

I have a geom_line plot in which one of the legends represents genotype. By convention in our field, one of these (the control) should be in plain font, while others (experimental knockout genotype(s)) should be italicized.

I have factorized the genotypes and tried

legend.text = element_text(face = c("plain","italic"))

but this doesn't make a difference.

Here is an example, with some quick dummy data

timecourse <- data.frame(rbind(abs(rnorm(10)),abs(rnorm(10))))
names(timecourse) <- prettyNum(10*abs(rnorm(10)),digits=3)
timecourse$Genotype <- as.factor(c("ControlGenotype","GeneX-/-"))
timecourse <- reshape2:::melt.data.frame(timecourse, id.vars="Genotype",variable.name="Timept", value.name="Y")
p <- ggplot(data=timecourse, aes(x=Timept, y=Y, color=Genotype, group=Genotype))
p <- p + geom_line(aes(), size=0.7) + geom_point(aes(), size=2.5) + scale_color_manual(values=c("#000000", "blue"))
p <- p + labs(x="Time (hrs)", y="Readout of some bioassay")
p <- p + theme (legend.text = element_text(face = c("plain","italic")))

"ControlGenotype" should be plain

"GeneX-/-" should be italicized (but is not)

Thanks very much,

Karolina

--

Karolina Maciag
MD/PhD Candidate, Harvard Medical School/MIT-HST
Hacohen Lab, Broad Institute

Dennis Murphy

unread,

Jul 8, 2015, 4:37:19 PM7/8/15

to Karolina Maciag, ggplot2

Hi:

Try this:

labl <- list("ControlGenotype", expression(italic("GeneX-/-")))

library(ggplot2)
ggplot(data=timecourse, aes(x=Timept, y=Y, color=Genotype, group=Genotype)) +
geom_line(size=0.7) + geom_point(size=2.5) +
scale_color_manual(values=c("#000000", "blue"),
labels = labl) +

labs(x="Time (hrs)", y="Readout of some bioassay")

Dennis

> --
> --
> You received this message because you are subscribed to the ggplot2 mailing
> list.
> Please provide a reproducible example:
> https://github.com/hadley/devtools/wiki/Reproducibility
>
> To post: email ggp...@googlegroups.com
> To unsubscribe: email ggplot2+u...@googlegroups.com
> More options: http://groups.google.com/group/ggplot2
>
> ---
> You received this message because you are subscribed to the Google Groups
> "ggplot2" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to ggplot2+u...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

Brian

unread,

Jul 8, 2015, 5:12:42 PM7/8/15

to Dennis Murphy, Karolina Maciag, ggplot2

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Beat me to it.

Why does changing

labl <- list("ControlGenotype", expression(italic("GeneX-/-")))

to
labl <- c("ControlGenotype", expression(italic("GeneX-/-")))
alter the justification?

So:

labl <- c("ControlGenotype", expression(italic("GeneX-/-")))

ggplot(data=timecourse, aes(x=Timept, y=Y, color=Genotype,
group=Genotype)) +
geom_line(size=0.7) + geom_point(size=2.5) +
scale_color_manual(values=c("#000000", "blue"),
labels = labl) +
labs(x="Time (hrs)", y="Readout of some bioassay")

Thanks
Brian

Karolina Maciag

unread,

Jul 17, 2015, 1:22:43 PM7/17/15

to Brian, Dennis Murphy, ggplot2

Thank you both - quite belated but I am really grateful.

It was a huge help - affects nearly every figure in my paper (and likely future papers as well), unless journal standards force the use of image editing software for legends.

I finally figured out on my own how to include Greek characters and newlines both in facet labels... everyone else I'd talked to about that had given up and used Illustrator!

More and more people in our lab are moving over to R for data exploration and figure generation, but we will need to make a handy cheat-sheet for the formatting in order to take the sweat out of annotating graphs.