kdetrees: Identify discordant trees in a sample

In grady/kdetrees: Nonparametric method for identifying discordant phylogenetic trees

Description

Analyze a set of phylogenetic trees and attempt to identify trees which are significantly discordant with other trees in the sample (outlier trees).

Usage

kdetrees(trees, k = 1.5, distance = c("geodesic", "dissimilarity"),
  outgroup = NULL, topo.only = FALSE, bw = list(), greedy = FALSE, ...)

Arguments

trees	multiPhylo object
k	IQR multiplier for outlier detection
distance	Select "geodesic" or "dissimilarity" distance calculation method
outgroup	if a character, reroot all trees with this species as outgroup. The geodesic distance method requires rooted trees.
topo.only	set all branch lengths to 1 before analyzing?
bw	see Details
greedy	greedy outlier detection?
...	additional arguments for distance calculation function, see details

Details

If bw is a single number, it will be used as a single constant bandwidth. It can also be a vector, in which case it will be used as variable bandwidths for each tree, repectively. Finally, if it is a list (default), the list will be passed as arguments to the bw.nn adaptive bandwith function.

... Is passed to either distory::dist.multiPhylo or dist.diss, as appropriate. See the help for these functions for more details.

Value

a kdetrees object; list(density,outliers)

Author(s)

Grady Weyenberg

Examples

kdeobj <- kdetrees(apicomplexa, outgroup="Tt")
print(kdeobj)
kdeobj$outliers

kdetrees(apicomplexa, k=2.0, distance="dissimilarity",topo.only=FALSE)

grady/kdetrees documentation built on May 17, 2019, 8:01 a.m.