RSEARCH version 1.0

Installation instructions for RSEARCH are in INSTALL.  The license is
in LICENSE.

The command syntax for RSEARCH is as follows:
    rsearch [options] <query> <target>

<query> is the structured RNA query file, in Stockholm format
(described in detail at 
	   http://www.cgr.ki.se/cgr/groups/sonnhammer/Stockholm.html 
Secondary structure base pairs are represented using '>' for the
opening nucleotide and '<' for the closing nucleotide.
Example query sequences are given in the examples subdirectory.

<target> is the target database in FASTA format.

The following options are commonly used:
    -h	      list all available options
    -c	      also search reverse complement strand
    -n N      Take N random sequences to determine extreme value
	      distribution paramaters.  (I usually use -n 1000)
    -s N      Use cutoff score of N (default 0)
    -E N      Use cutoff e-value of N (default is to use score cutoff)
    --noalign Don't report alignments between query and target.  This
	      can save time, especially when 
    --partition N[,M]...     Use partition points of N,M,... to
			     get different values for lambda and K
			     based on G+C content

If you are using the MPI version, the command is
mpirsearch [options] <query> <target>

You of course need to run this binary using "mpirun."  For instance, 
running under the LAM implementation of MPI might look like:

mpirun C mpirsearch query.stk target.fa

The most common command I use would be something like

rsearch -n 1000 -c -E 10 query.stk database

to search the RNA sequence in query.stk against both strands of the
database file, using an E-value cutoff of 10.

For each hit that's reported, the sequence name is prepended with > at the
beginning of the line for easy grepping.  Positions in the query and target
are reported, along with scores and (if appropriate) E-values.  

Robert J. Klein
rjklein@genetics.wustl.edu
April 8, 2003