Palign, pmembr etc

INTRODUCTION

This is a set of programs for sequence searches. It can perform both
dynamic programming and heuristic searches. Unfortunately they are not
very well documented


LICENSE


This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; version 2 of the License. With the
exception that if you use this program in any scientific work you have
to explicitly state that you have used PALIGN and cite the relevant
publication (dependent on what you have used PALIGN for). My
publication list can be found at http://www.sbc.su.se/~arne/papers/.
This program is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
General Public License for more details. You should have received a
copy of the GNU General Public License along with this program (in 
the file gpl.txt); if not, write to the Free Software Foundation, 
Inc., 59 Temple Place - Suite 330, Boston, MA 02111-13

For support please use email to arne@sbc.su.se and add PALIGN-SUPPORT 
in the header

BINARIES:

In the tar distribution there are binaries for linux (redhat 7.2)
compiled in linux/p*. Note that these are compiled for quite large
sequences and therefore you need about 700mb of RAM+SWAP space.



COMPILATION

You might have to define the environment variable ARCH (either in your
shell or in Makefile) and maybe edit the file $ARCH/flags.mk (not if
you are using g77) that set some flags that are specific for your
computer. Then all you should need to do is to type:

> make all

this will place executables in $ARCH/ 

(Note that precompiled binaries for redhat 7.2 exists in the linux/
subdirectory). Development is done using gcc and intels fortran
compiler. If you have this the code will be significantly faster
(binaries in linux.ifc/). However, porting should be quite trivial to
any plattform.

PMEMBR

To use Pmembr (Hedman et al, 2002) you also need tmHMM (available from
Anders Krogh) or another membrane prediction program (that can produce
a similar output). You will also need to convert the output of tmHMM
to our database format using the plptofast.pl program.

To run pmembr you need to (a) compile the program (b) run tmHMM for
your sequence and your database (c) convert your database to our
internal format with plptofast.pl. {The database can also be download
from http://www.sbc.su.se/~arne/pmembr/sprot.seq} (d) run PSIBLAST
(with these parameters

> blastpgp -j 5 -h 1e-5 -m 1 -e 99 -v 10000 -F F against
SwissProt) and (e) run it with this command.  

pmembr2 -prof FILE.plp FILE.psi - DATABASE 




DIFFERENT PROGRAMS: 

There are three groups of programs:
- palign, palign2 and palignh are for searches of globular proteins
- pmembr, pmembr2 and pmembrh are  for searches against membrane proteins 
- palignp and palignp2 are for profile-profile alignments

in each group there are two or three program:
- pmembr, palignp and palign just aligns two sequences
- palign2,palignp2 and pmembr2 performs a search of a sequence (or profile)
  against a database of sequence.
- pmembrh/palignh are heuristic versions of pmembr2 and palign2 using a
		FASTA-like algorithm for the algorithm. Not perfect
		but they work and are quite fast. These are not maintained any more.



USAGE:
type palign -h for some help on different flags that can be used.
Palign can be used for many type of alignment problems, it is
certainly not always the fastest of all programs but it is very flexible 
and can:

- Align two sequences and search a database with a sequence
- Align a sequence and a profile
- Include and combine secondary structure information in the alignments
- Read sequences, profiles etc in many formats
- Perform profile-profile searches/alignments using different algorithms

EXAMPLES:

Alignments:
- To align two sequences (and print the alignments)
> palign -ali seq1.seq seq2.seq
- To align a sequenc againast a pdb file and print the coordinates of the model into file foo.pdb
> palign -pdb foo.pdb seq1.seq seq2.pdb
- To use secondary structure information in the alignments
> palign -seq seq1.seq seq1.ss - -prof seq2.seq seq2.ss -
- Profile files can be read from PSIBLAST (and some other less used formats)
- to obtaine them use the -Q switch in PSIBLAST
- To align two profiles using the log average scoring
> palignp -logaver -prof prof1.psi- -seq prof2.psi -
- To use the ProfNet method and print the coordinates of the model into file foo.pdb.
- -scalepsib transform the values in the log-odds based psi-blast profile vector.
- -winsize and -win should be set to 1 and -lambda should be set to 2 to use the correct scoring algorithm.
- Here the GO/GE and shift values are set to 0.3 0.03 -0.3
- NB the order of the arguments is very important
> palignp -lambda 2 -scalepsib -winsize 1 -win 1 -go 0.3 -ge 0.03 -shift -0.3 -readnn profnet_nin40_nhid29_out1_epochs700.ann -pdbfile model_of_file1.pdb -prof file2.pdb file2.psi - -seq file1.psi -

Searches:
- To search a sequence against a database (list of sequence files)
> palign2 seq1.seq list.txt
- To do a profile--profile  search 
> palignp2 seq1.psi list-psi.txt



PLATFORMS: 

The programs are normally developed on linux using either g77 or
intels compilers.  Some of the programs have been tested (compiled) on
alphas/aix/sgi/sun4 machines but this is not at all supported, but
should probably work. You need to either use the split.awk script or
have a compiler that (a) inludes cpp preprocessing and (b) lines
longer then 72 characters. The rest should be more or less standard
fortran 77 Look in */flags.mk for ideas on how to compile. Note that
the profile-profile comparisons are written in C.


TROUBLESHOOTING

1) Make sure you have the latest version

2) If you get segfaults it is possible that you do not have enough
   memory, try then to change maxseqlen and maxproflen in dynprog.h and
   recompile.
