Full
model name: Any of the reversible 4x4 models.
Model type:
Nucleotide
Model
options:
Custom models are defined by means of a 6 character
string which is is shorthand for defining 6 rate matrix elements
which are located above the diagonal, left to right, top to bottom.
The symbols in the string should be from the set '0','1','2','3','4','5'.
If the two symbols in the string are identical, then the two
corresponding entries of the rate matrix will be the same (see
example below). Each model has the following options:
- Local.
Rates are homogeneous (same
for all sites in the data). Each branch as many parameters as
there are different characters in the model string.
- Global.
Rates are homogeneous (same
for all sites in the data). Each branch one parameters, there
are several shared ratios for the tree (the number of ratios
is the same as the number of different characters in the model
string less one).
- Rate Variation.
Different sites may have different rates. The
user specifies how many different rate classes there should and
selects one of the available distributions
to be sampled. Distribution parameters are estimated.
- Rate Variation+HM. Similar to Rate Variation except that rates
at adjacent sites are correlated via a Hidden Markov Model
Independent
Parameters:
Will depend on
the model string. See the following Example.
Example:
Model string 012134
Rate Matrix:
- Local.
5 independent parameters per branch:
a,b,c,d,e.
- Rate Variation, Rate Variation+HM:
1 independent parameter per branch:
t.
4 shared ratios: B,C,D,E.
with the vector representing
equilibrium nucleotide frequencies. '*' is defined to be the
negative of the sum of all other row elements.
Equilibrium
frequencies. There are 3 options
to choose from:
- Observed: The proportions of nucleotides from the data
will be used.
- Equal:
All frequencies are set to .25.
- Custom:
The user specifies the frequencies for each nucleotide.
|