Cluttered/obscured node pie charts

[Gaby]

Hi, I'm a new user exploring BioGeoBEARS for my PhD research on a genus of neotropical birds. I've run the basic analysis script, but I'm running into an issue with the graphical results: the pie charts on many nodes are completely obscured and impossible to interpret because the slices are way too small. I suspect this is due to high ambiguity in the ancestral state reconstruction.

I assume the tiny slices represent states with very low probability. Is the best way to address this to add more biologically realistic parameters to the model to reduce this ambiguity?

Given that my study group are not strong flyers, so maybe I could try adding a distance matrix.

Also, access to areas A and B (North of the Panama Isthmus) was severely limited until around 5 Mya. So maybe I could implement this as a parameter too?

I'm no sure if any of this would be enough to improve the clarity of results, or if it's just due to the characteristics of my system.

 I am still working through the documentation and tutorials, so it's very possible I've missed something fundamental. Any guidance on which of these paths is most likely to help, or any other strategies to improve the interpretability of my results, would be greatly appreciated.

Thank you for your time.

Best,

Gaby

[Nick Matzke]

Hi -- you can change the size of the PDF, or of the pie charts.  But that won't help with the "so many slices it's looking black" issue.

As you say, that genuinely reflects the uncertainty of the estimate for those nodes. That's a result, conditional on the model and data and tree, the only thing that will change it is changing the model or data (or tree).  E.g. it could be that you have over-split the areas given your number of species.  If a bunch of species are widespread, it's common to have uncertainty, as there are many ways to have multi-area ranges, and many ways to convert between them.  But if you collapse it to (e.g.) just North and South, or some such, you might get more clarity for that coarse question.

The other option is a per-area probabilities plot, this can be useful in these situations:

http://phylo.wikidot.com/example-biogeobears-scripts#per-area_probabilities

http://phylo.wikidot.com/example-biogeobears-scripts#per-area_probabilities_graphics

Cheers,

Nick

--
You received this message because you are subscribed to the Google Groups "BioGeoBEARS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to biogeobears...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/biogeobears/65e0e334-8d17-4e25-a5b4-54cef52a1cd4n%40googlegroups.com.