Previous: CountClusters, Up: Descriptive [Contents]
ModelFit fits a substitution model to the input block, given a phylogenetic tree.
All nucleotide homogeneous models can be used, with or without rate across sites variation.
maf.filter= \
[...],
SequenceStatistics( \
statistics=(\ \
[...],
ModelFit( \
model=HKY85( \
kappa=1, \
theta=0.5, \
theta1=0.5, \
theta2=0.5), \
rate_distribution=Gamma( \
n=4, \
alpha=0.5), \
tree=BioNJ, \
parameters_output=( \
HKY85.theta, \
HKY85.theta1, \
HKY85.theta2, \
HKY85.kappa), \
fixed_parameters=(), \
reestimate_brlen=yes, \
max_freq_gaps=0.3, \
gaps_as_unresolved=yes)), \
[...]), \
ref_species=species1, \
file=data.statistics.csv), \
[...]
|
model={string}Substitution model to use. See the Bio++ Program Suite manual for an extensive description of available models. All nucleotide models can be used.
rate_distribution={string}The distribution for rates across sites. See the Bio++ Program Suite manual for all available distributions.
root_freq={None|Full|GC}Allow root frequencies to be different (non-stationary model). Root frequencies can be fully parametrized, or parametrized with GC content.
tree={string|none}The property name under which trees are stored for each block. If set to “none”, then an input file should be given.
tree.file={path}[tree=none]Path for tree file, in case no property is set.
tree.format={Newick|Nhx}[tree=none]Format for tree file, in case no property is set.
parameters_output={list}A list of parameter names to output as statistics.
fixed_parameters={list}A list of parameters which should not be optimized, but fixed to their initila values.
reestimate_brlen={boolean}Tell if the branches of the tree should be reestimated for each block together with other model parameters.
max_freq_gaps={float}The maximum proportion of gaps for a site to be included in the analysis.
gaps_as_unresolved={yes/no}Tell if remaining gaps should be converted to ’N’ before likelihood computation. This should be ’yes’ unless you specify a substitution model which explicitely allows for gaps.
global_clock={yes/no}Assume a global clock for branch lengths.
reparametrize={yes/no}Transform parameters to remove constraints (can improve optimization, but is usually slower).