It is possible to reduce the parameter space by putting extra constraints on parameters, using for instance
model=TN93(kappa1=1.0, kappa2=kappa1, theta=0.5)
In that particular case the resulting model is strictly equivalent to the HKY85 model. This syntax however allows to define a larger set of models.
As long as their range match, parameters of several objects (models, root frequencies, rates, etc) can be linked.
For instance:
model1 = T92(theta=GC.theta, kappa=3) model2 = T92(theta=0.39, kappa=T92.kappa_1) nonhomogeneous.root_freq=GC
In the case of nonhomogeneous modelling, a specific syntax is available:
nonhomogeneous.alias = {list of aliases}
where each alias is described as ‘param1->param2’. The full name of the parameters have to be used, see for example:
model1 = T92(theta=0.4, kappa=4) model2 = GTR(theta=0.4, a = 1.1, b=0.4, c=0.4, d=0.25, e=0.1) nonhomogeneous.alias=GTR.theta1->T92.theta1
This option can be used to link parameters of the root frequencies if the model is non-stationary:
model1=GTR(theta1=0.7) nonhomogeneous.root_freq=Full(init=balanced) nonhomogeneous.alias=Full.theta1->GTR.theta1_1
Note that this option is only available with the ’general’ nonhomogeneous substitution models and will be ignored if used with "one_per_branch".