Optimizing Substution Matrix by Separating Score Distributions

Optimizing Substitution Matrix by Separating Score Distributions

Paper

[Download PDF from OUP] (Personal research/study purposes only.)

Download Source

The program package used for our experiments is available from here.
The previous version of this package is here(Jul 11 2002).
This software requires only a standard unix platform.

Compile

First, extract the source.

$ gzip -cd optssd-3.0.1.tar.gz | tar xvf -

To compile this program, go to the extracted directory

$ cd optssd-3.0.1

and type as follows:

$ ./configure

Then, edit src/Makefile.common if required.
To compile the source, type

$ make

Load modules will be in 'bin' directory.

Usage

To execute optssd, one must prepare

a setting file for optimization and score distribution,
a setting file for an alignment method,
a classified database, and
a substitution matrix.

Optssd Setting File1

Default filename is input.prj.
An example is here. The format will be clear.

Optssd Setting File2

Second file must be pointed by setting UDOptions in the file above.
An example is here.

Some score parameter files (substitution matrices) are in 'scr' directory.
And some databases are also available from here.

Then, type the command,

$ bin/optssd [input.prj]

Output Files

optssd.log	Experimental environment.
itnnn.scr	A score function on nnnth optimization step. It can be converted to the popular format by "`scutil -B`" command.
scr.log	The log of objective function and the minimum number of errors
*.as	scores of alignments
cgdata.*.log	EVD parameters for each group

Databases and Optimized Matrices

Database	Classified Sequence Files	Group File	Optimized Matrix
COG	COGs.tar.gz	COG.grp	COGOPT.mat
Pfam	Pfam.tar.gz	Pfam.grp	PFAMOPT.mat
SCOP 40%ID	scop1.61-40.tar.gz	scop1.61-40.grp	SCOP40OPT.mat
SCOP 95%ID	scop1.61-95.tar.gz	scop1.61-95.grp	SCOP95OPT.mat

Links

hourai@is.s.u-tokyo.ac.jp