phangorn (version 2.5.5)

SH.test: Shimodaira-Hasegawa Test

Description

This function computes the Shimodaira--Hasegawa test for a set of trees.

Usage

SH.test(..., B = 10000, data = NULL, weight = NULL)

Arguments

...

either a series of objects of class "pml" separated by commas, a list containing such objects or an object of class "pmlPart" or a matrix containing the site-wise likelihoods in columns.

B

the number of bootstrap replicates.

data

an object of class "phyDat".

weight

if a matrix with site (log-)likelihoods is is supplied an optional vector containing the number of occurances of each site pattern.

Value

a numeric vector with the P-value associated with each tree given in ....

References

Shimodaira, H. and Hasegawa, M. (1999) Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Molecular Biology and Evolution, 16, 1114--1116.

See Also

pml, pmlPart, pmlCluster, SOWH.test

Examples

# NOT RUN {
data(Laurasiatherian)
dm <- dist.logDet(Laurasiatherian)
tree1 <- NJ(dm)
tree2 <- unroot(upgma(dm))
fit1 <- pml(tree1, Laurasiatherian)
fit2 <- pml(tree2, Laurasiatherian)
fit1 <- optim.pml(fit1) # optimize edge weights
fit2 <- optim.pml(fit2)
# with pml objects as input
SH.test(fit1, fit2, B=1000)
# in real analysis use larger B, e.g. 10000

# with matrix as input
X <- matrix(c(fit1$siteLik, fit2$siteLik), ncol=2)
SH.test(X, weight=attr(Laurasiatherian, "weight"), B=1000)
# }
# NOT RUN {
example(pmlPart)
SH.test(sp, B=1000)
# }

Run the code above in your browser using DataCamp Workspace