RNAsifter - a shape based filter for Rfam
http://www.bibiserv.uni-bielefeld.de/RNAsifter
----------------------------------------------------------------

:: About this software:

   RNAsifter is a non-rigorous filter for the rfam database, 
   that is based on the abstract shape approach. It exploits 
   the familily specific secondary structure, i.e. it computes 
   shapes for the query-sequence and compares them to a 
   precalculated shape-index. The exact matches deliver - hope-
   fully only few - candiate families in which the query-
   sequence might fold. Membership for the candidate families 
   is checked with the program cmsearch, afterwards.
   
   
:: Getting RNAsifter:
   
   WWW:	http://www.bibiserv.uni-bielefeld.de/rnasifter
   
   
:: Installation:
   
   RNAsifter is a simple Perl script. So you need a current
   version of perl. We used perl-5.8, but it should also work 
   with other versions.
   
   RNAsifter depends von RNAshapes, to calculate abstract shapes
   of the query. You have to install this program locally to 
   execute RNAsifter. RNAshapes is available for download at
	http://www.bibiserv.uni-bielefeld.de/rnashapes
   
   RNAsifter also depends on the program cmsearch, a part of 
   the infernal package. As before you have to install this 
   software first. Currently RNAsifter works with version 0.71, 0.72
   and 0.81.
	http://infernal.janelia.org/
   
   As you are reading this file, you successfully unpacked the
   tar.gz archive. This archive contains the two Perl scripts:
   RNAsifter.pl is the main program, that searches queries against
   a database, e.g. Rfam. RNAsifterBuild.pl is for constructing your
   own shapeindices, if you have to search a new database or decide
   to run with more shapes per query. We suggest to use our pre-
   calculated shapeindices, because a RNAsifterBuild run for Rfam 
   takes several _days_ on a standard machine.
   So to run RNAsifter you must download one of our precalculated
   shapeindices: http://bibitest/download/tools/rnasifter.html
   These packages contains all needed cm-models 
   (directory models/), the score thresholdfile Rfam.thr and
   our previously calculated Shape-Index for Rfam.

   
:: Getting started:
   
   A typical command line call could be like this:
   perl RNAsifter --d models/ --thresholdfilename
   
   perl RNAsifter.pl --cmsearchBinDir /usr/bin/cmsearch --RNAshapesBinDir /usr/bin/RNAshapes --d models/ --thresholdFile Rfam.thr ShapeIndex__Rfam8.0__full__5-5-5-5-5.store example.fasta
   
   The two parameters --cmsearchBinDir and --RNAshapesBinDir
   are unnecessary if you point to the correct directories with
   your PATH variable. You can also set system variables
   INFERNAL_BIN_DIR and RNASHAPES_BIN_DIR.
   The system variable RFAM_DIR frees you from setting the 
   --d parameter.
   
   The output should look like this:
   
   AY362699.1/13399-13480_RF00507       1      82   RF00507       1      60     70.75      Corona_FSE
   

:: ShapeIndexBuild.pl
   
   The default Shape-Index has been calculated for Rfam 8.0,
   up to the first five Shapes for each sequences of the full
   alignments. Perhaps you might want to do your own calculations,
   maybe with more shapes, use the program ShapeIndexBuild.
   Note: a complete run for Rfam 8.0 may take several days!!
	   
