- Use then the Shimodaira–Hasegawa test [44] implemented
into RAxML to test whether the ML tree explains the data
significantly better than the alternative tree, in which the fungi
are placed such that they agree with the species phylogeny. To
do so runraxml –n outputFileName –s alignmentFilename –m
modelOfChoice –f h –t mlTreeFilename –z alternativeTreeFile-
name. The program will then tell you whether or not the ML
tree is significantly better than the alternative tree. In the
example of Fig.11, the sequence tree modified to reflect the
evolutionary relationships of the species is only 3.9 log likeli-
hood units worse than the ML tree. According to the SH test,
this difference is not significant (p<0.01), and thus both trees
explain the data equally well. It is therefore unnecessary to
invoke complex evolutionary scenarios (seeFig. 10d) to explain
the deviations of the ML tree from the species tree topology.
0.5
PRKAG1-3 Arabidopsis thaliana
KING1 Arabidopsis thaliana
PRKAG1-3 Sorangium cellulosum
SDS23 Candida parapsilosis
PRKAG1-3 Nematostella vectensis
KING1 Oryza sativa
SDS23 Saccharomyces cerevisiae
KING1 Selaginella moellendorffii
KING1 Capsaspora owczarzaki
PRKAG1 Homo sapiens
PRKAG1-3 Drosophila melanogaster
KING1 Dictyostelium discoideum
SDS23 Spizellomyces punctatus
PRKAG3 Homo sapiens
PRKAG1-3 Ectocarpus siliculosus
PRKAG1-3 Pyrobaculum neutrophilum
PRKAG1-3 Chlamydomonas reinhardtii
PRKAG2 Homo sapiens
SDS23 Kluyveromyces lactis
PRKAG1-3 Saccharomyces cerevisiae
PRKAG1-3 Selaginella moellendorffii
PRKAG1-3 Batrachochytrium dendrobatidis
KING1 Polysphondylium pallidum
PRKAG1-3 Phytophthora sojae
PRKAG1-3 Ustilago maydis
0.5
Species tree topology
Maximum likelihood tree topology
LogLik = -16,307.6 LogLik = -16,311.5
40
99
54
76
84
66
92
100
100
95
86
100
62
82
60
78
81
99
90
100
51
76
*
* *
Fig. 11The evolutionary histories of the human AMPKγsubunit and of its homologs in plants, AMPKβγand
KING1. The tree to the left represents the maximum likelihood (ML) topology computed with RAxML
(substitution model: LG [42]þIþG). Branch labels denote percent bootstrap support. Asterisks indicate
gene duplication events. The subtree harboring AMPKγof animals together with its fungal and plant orthologs
is shown in black. A gene duplication event in the last common ancestor of the eukaryotes gave rise to the
gray subtree, which harbors KING1 and its orthologs. This rejects the hypothesis that KING1 is the plant
counterpart to animal AMPKγ(see[12] for details). The positioning of the fungal sequences in both trees
(hatched clades) is at odds with the placement of fungi in the eukaryote species phylogeny (right tree). The log
likelihoods of the data given the ML tree and given a tree placing the fungal sequences at the expected
position are given below the respective trees.Seethe main text on tree topology testing for further details
Tracing AMPK Evolution 135