Biology Reference
In-Depth Information
Table 1
Relationship between algorithmic parameters and software settings
Algorithmic
parameter
Software setting
choices
Software setting
Description
Subproblem
alignment
method
“Aligner” dropbox
MAFFT, ClustalW,
Prank [
26
],
Opal [
27
]
This determines the method used
to align the subsets of the
sequences, and the default is
MAFFT
Alignment
merge
method
“Merger” dropbox
Muscle, Opal
This determines how
subset alignments are merged
together. Muscle is the default,
but Opal should be used if the
dataset is small enough
ML-based
phylogenetic
estimation
method
“Tree Estimator”
dropbox
FastTree, RAxML
This determines how
phylogenetic trees are
estimated in each iteration; the
default is FastTree, due to its
improved speed relative to
RAxML
Substitution
model
“Model” dropbox
Many models,
depending on type
of data and ML tree
estimation method
The model selected determines
the parameters optimized by
the ML tree estimation
method. See [
29
]
Maximum
subproblem
size
“Max. Subproblem”
dialog
Percentage (1-50 %),
size (1-200)
This determines the maximum
size subset given to the
“aligner” method
Decomposition
edge
“Decomposition”
dropbox
Centroid, longest
This determines the edge used to
decompose the dataset into
subsets. The default setting is
the centroid edge
Termination
condition
“Apply Stop Rule”
dropbox
After Last
Improvement, After
Launch
This determines when the
stopping rule is evaluated. We
recommend using “After Last
Improvement” unless your
dataset is very large
Termination
condition
“Stopping Rule” dialog -
“Blind Mode
Enabled” checkbox
Checked/unchecked
This determines which tree (best
ML or current tree) is used in
the subsequent iteration
Termination
condition
“Stopping Rule” dialog -
“Time Limit (hr)”
dialog/”Iteration
Limit” dialog
0.01-72 h (“Time Limit
(hr)” dialog)
1+ (“Iteration Limit”
dialog)
This determines whether time or
number of iterations is used to
define when SAT´ stops
Final tree/
alignment
pair output
by SAT´
“Return” dropbox
Best, Final
This determines which tree and
alignment pair (Best ML or last
pair computed) is output
(continued)
