Closed CodeInTheSkies closed 4 years ago
Hey, it's certainly possible to do this but perhaps not with EnhancedVolcano as it's currently designed. You could generate two separate plots and place them side-by-side via _cowplot::plotgrid or using the grid and gridExtra packages, or else try to come up with some other cutomised solution.
Thanks, Kevin. Yes, I am aware of these possibilities. I am looking to see if I can code this somehow, but in case you have some ideas about how to specifically achieve this, i.e., having positive values on the left of zero as well, and also changing the colors for left and right separately, please let me know, if you wouldn't mind. Thanks.
As it turns out, you can sort of 'cheat' to do this via scale_x_continuous()
. If you take a look at this example from the vignette: https://github.com/kevinblighe/EnhancedVolcano#custom-axis-tick-marks
Now, we can simply change the labels via:
p1 +
ggplot2::coord_cartesian(xlim=c(-6, 6)) +
ggplot2::scale_x_continuous(
breaks=seq(-6,6, 1), labels = c(6,5,4,3,2,1,0,1,2,3,4,5,6))
So, that should do it. The underlying values are still negative on the left-hand-side, though, but you can deal with this, correct?
Thanks a lot, Kevin! That trick works wonderfully!
Excellent!
Hi Kevin,
Thanks for writing this very useful EnhancedVolcano package! Lovely!
I have a rather unusual customization scenario, where I have two conditions, say WT and KO, each with their own lists of LogFC and adj_pVals. I am only interested in the positive LogFCs. But I wish to plot both the conditions in a single plot, where the left of zero would correspond to WT cases (with positive x-axis values), and the right would correspond to the KO cases, again with positive values. Of course, as per your Vignette, I think I can also choose different colors for the two conditions.
Essentially, since I am not plotting negative logFCs at all, I don't need the negative x-axis part, but instead would like to replace that with the second condition.
I know I can do this another way, where I limit the x-axis to only the positive values, and then plot each condition with a different color, but the problem would be that the points would overlap a lot as they would have similar LogFC values and pValues. So, it would be best if I can split as left and right for the two conditions, so that I can neatly label the genes too without too much overcrowding.
Do you think this is possible?
Would very much appreciate your insights!
Many thanks.