Add BEAST dataset

[thibautjombart]

From one of the well-known practicals. We basically just need to posterior sample of trees

[MichelleKendall]

I have four sets of BEAST trees (with four different sets of priors) for dengue fever, built from the xmls supplied in the original BEAST paper http://www.biomedcentral.com/1471-2148/7/214/additional

Is that the sort of thing you're thinking of?

[thibautjombart]

Yes absolutely!
I have added a HA H3N2 dataset they used in a tutorial at 5167cf2
(this is on branch newfeatures)

BTW, data can be documented using roxygen2 too. I made that change - see https://github.com/thibautjombart/treescape/blob/newfeatures/R/data.R

[MichelleKendall]

Sorry I completely forgot this was on branch newfeatures. I've now included the flu trees and proper documentation in the master branch, along with the Dengue trees as promised. I haven't put the flu trees as an automatic example in Shiny because they're quite big and a bit slow, but they're a great example to include in the package, thanks.

You said you used the first 200 trees from the BEAST run - did you already discard some trees as a burn-in? It usually makes more sense to use the final 200 or a random sample of 200 from the second half, was there a reason not to do that here? Thanks.