Sunday, March 13, 2016

Fixes to dotTree & fitPagel; update to consensus.edges

Yesterday & this morning I pushed a couple of new phytools updates to GitHub (1, 2).

Among the two updates, (1) I fixed a bug in fitPagel for cases in which not all pairwise states of the two binary characters (0|0, 0|1, 1|0, and 1|1) are represented in the dataset. I fixed this bug for method="fitMk" (the default), but it may persist for the other two model-fitting methods. I also fixed a bug in dotTree for data.type="discrete" that was producing a really weird result in terms of the color scheme and legend.

(2) I made a small update for consensus.edges so that instead of checking if the tree is rooted for method="least.squares", the function checks if the trees are all ultrametric. This is done because the internally used function, nnls.tree(...,rooted=TRUE), will return ultrametric branches - which does not make sense if the original trees were not ultrametric!

That's it. These updates can be obtained by installing phytools from GitHub as follows:

library(devtools)
install_github("liamrevell/phytools")
library(phytools)

1 comment:

  1. Amazing timing. This is just what I was hoping for.
    Thank you!

    ReplyDelete