78
8. Assess the robustness of the ML tree to variations in the substitution model by repeating the tree inference in PhyML
using alternative substitution matrices that scored highly in
Subheading 3.3. The resulting trees should be very similar to
the maximum- likelihood tree.
1. Ensure that the alignment file and tree file are properly formatted for input into PAML [26] (see Note 13).
2. Set up the program by modifying the control file (“codeml.ctl”),
as shown in Table 1.
3. Run the program from the command line using the command
codeml.
3.5 Reconstruction
of Ancestral
Sequences
Table 1
Variables in the control file for ancestral protein reconstruction
using PAML
seqfile
Filename of alignment
treefile
Filename of tree
outfile
Filename of output
noisy
9
verbose
2
runmode
0
seqtype
2
clock
0
aaRatefile
Filename of rate matrix (e.g., lg.dat for the LG matrix)
model
2
Mgene
0
fix_alpha
0 to use +G model, else 1
alpha
0.5 to use +G model, else 0
Malpha
0
ncatG
Number of rate categories to use for the +G model
getSE
0
RateAncestor
1
Small_Diff
.5e-6
cleandata
0
method
0
Deviations from the settings in the default control file are indicated in bold. The
remaining variables in the control file are not applicable to amino acid-based analysis
and can be removed
Ben E. Clifton et al.