Suggestions for plotting options

[mbutler808]

Some suggestions for plotting options that are better for large phylogenies or long taxon labels. We had a problem with not being able to read the metadata because it was cut off

https://github.com/kingaa/ouch/blob/master/R/plot.R#L17
Add frame.plot, cex, offset, rtspace arguments for tree.plot.internal:

# most users will not want the frame around the plot, but specifying it as part of ... generates a warning because itʻs not a valid argument to text() 
# the offset and rtspace is set to allow more space for taxon names (and at least show users what to change if need be)
# also changed the figure size to accommodate long taxon names
tree.plot.internal <- function (x, regimes = NULL, labels = x@nodelabels, legend = TRUE, frame.plot=F, cex=1, offset=0.1, rtspace=1, ...) {

More room for figure and Check size allotted to labels by character length:

    par(yaxt='n', mar=c(4, 1, 1, 1))
    plot(X,Y,type='l',col=C,xlab='time',ylab='',xlim=rx+c(-rxd,rtspace*rxd+max(strwidth(labels, units="figure"))),ylim=c(0,1), frame.plot=frame.plot,...)

And did you mean to include ... for text or only for plot? But if it is not a valid option it will throw a warning. The cex and offset are used here.

     text(X[seq(1,length(X),6)],Y[seq(1,length(Y),6)],labels[f],pos=4, cex=cex, offset=offset)

Also it would be really nice to be able to specify the pallette as an argument, but I am mulling over the best way to associate it with regime. - attach a regime name to the value?

These are small things, but really annoying when youʻre trying to get the figure right for publication.
Thank you!
Marguerite

plot.R.txt

[kingaa]

@mbutler808: some of these things are easily accommodated. Others require substantial thinking. On the whole, I'm wary of including all the options present in lower-level functions, because this becomes a headache to maintain. I agree that it is unfortunate that the ... is passed both to text and to plot. This is clearly a problem. I suppose one way to deal with it is to have ... passed, e.g., to plot and another option, say text_opts passed to text. There is also legend....

More generally, preparation of publication-quality figures need not be a goal of ouch, especially in the presence of other packages that have this as a goal. For example, I am very fond of ggtree, but I suppose there are others. The goal of ouch should be to support analysis using the models it provides. This does include a certain amount of graphics, so that users can e.g., load data and verify that it is entered correctly, paint a tree and check that it is done properly, visualize alternative hypotheses, etc. However, preparation of publication-quality graphics is something better left to other packages.

[kingaa]

Just pushed a new branch. Please have a look. Access to the options of the low-level functions plot, text, and legend is now possible. Ladderization is an option, with the default being to ladderize. The color palette is customizable.

[mbutler808]

Thank you so much Aaron. Really appreciate you working on this. I checked it out and marked a couple of things in new issues. I understand the pain in terms of maintenance, the publication-quality remark was just a bonus. Seems to me that this is very close to publication quality, and a lot of people would just go with this.

But the more important issue is plotting the trees and regimes for checking your data analysis. I should have indicated which was the biggest problem -- cutting off names. It has hindered us because we have a lot of unnamed species and long meta data. We used just short codes for a while, but finally had to check and there were errors that we missed because we couldnʻt see the names and/or sites. Itʻs an easy fix, I think, just need to calculate the space based on character length which I provided. This is really important for the analysis set-up and checking. Once everything is correct, of course we would make shorter names for publication.

Also one other comment - ouchʻs plotting routines are SO MUCH FASTER than ape or phylotools. Itʻs noticeable. So much more pleasant to do data exploration. My student is doing a biogeographic analysis on this huge phylogeny and needs to paint it in many different ways.

I have avoided ggplot and therefore ggtree, because ggplot is poorly documented and the syntax is so different. It was too frustrating to learn. But I may have to eventually.

Thanks,
Marguerite