Biology Reference
In-Depth Information
Several conventional consistency-based schemes have been
proposed in the literature [
10
-
13
]. These schemes are primarily
designed to overcome the drawback of progressive alignment
through evaluating the accuracy consistency between the resulting
MSA and pairwise alignments. ProbCons [
14
] introduced a new
probabilistic consistency model, which is used to reestimate more
accurate pairwise posterior probabilities by a three-way alignment
criterion. Different from the conventional schemes, this probabilis-
tic consistency model does not iteratively evaluate the accuracy
consistency between the resulting MSA and pairwise alignments.
Since MSAProbs relies on pairwise posterior probabilities to com-
pute MSA, it employs a similar probabilistic consistency model to
ProbCons with the difference that sequence weights are introduced
to the model to avoid a biased sampling of sequences. The trans-
formations are further performed for a fixed number of iterations to
refine the posterior probabilities. By default, our program uses two
iterations in order to offer a good trade-off between alignment
accuracy and speed.
3.4 Consistency-
Based Schemes
Before showing how to align sequences using MSAProbs, we list
the major command line parameters accepted by the software.
3.5
Program Usage
-num_threads
n: Specifies the number of threads to be used
for parallel execution. Otherwise, the program will automati-
cally detect the number of available CPU cores in the computer
and then use the same number of threads with the number of
CPU cores.
l
-c n: Specifies the number of iterations performed for weighted
probabilistic consistency transformation (default
l
¼
2).
-ir
n
:
Specifies the number of iterations performed for random-
split iterative refinements (default
l
¼
10).
-o
file: Specifies the alignment output file name
file
. By default,
the resulting alignments will be output to STDOUT.
l
-clustalw
: Indicates that the resulting alignments will be dis-
played in CLUSTAL format rather than the default FASTA
format.
l
-a
: Indicates that the aligned sequences will be displayed in
alignment order rather than the input order. Otherwise, they
will be displayed in the same order as in the input file.
l
Unlike other MSA algorithms requiring users to tune a number
of parameters in order to yield higher alignment accuracy, MSA-
Probs attempts to unload this tedious burden from users by fully
automatic execution while achieving high alignment accuracy.
However, we still provide two parameters to allow users to have
some control over the procedures relating to alignment accuracy.
