.TH "xblast" 1 "Jan 26, 1994"
\" This is the on-line manual page for xblast.
\"
\" The operation of the 'man' command can be simulated with the command:
\"		nroff -man xblast.1 | ul | more
\"
.UC 4
.SH NAME
.LP
.B xblast   
- read blast output and mask query

a companion program to the blast suite.
.SH SYNOPSIS
.LP
.B   xblast  <blast_output>  <query_filename>  [masking_char]
.LP
.B | xblast       +          <query_filename>  [masking_char]

.B   xblast  -d <blast_output>  <query_filename>  [masking_char]
.LP
.B | xblast  -d     +           <query_filename>  [masking_char]

.B   xblast  -r <blast_output>  <query_filename>  [masking_char]
.LP
.B | xblast  -r     +           <query_filename>  [masking_char]

.br
.SH DESCRIPTION
.LP
Functionally,
.I xblast
will read a blast.output file and generate a query sequence
where all segments with hits are masked.
It will read output from blastp/blastn/blastx/tblastn and mask
the corresponding nucleotide or amino-acid query.
.LP
The primary motivation for this program is to filter out unwanted hits.
xblast may be used in the following situations:
1) by first running the relevant blast search against a database of
"junk sequences" (like alu repeats) and then use the masked query for
the "real" search,
2) for masking already known similarities (e.g. with an ubiquitous domain)
and working on the rest of the sequence with increased sensitivity 
(e.g. PAM120 then PAM250).
.LP
xblast parses the human readable blast output. It looks for
the 
.B Query: [pos1] [sequence] [pos2] 
lines found after the 
.B Sequences producing High-scoring ....
line. Thus blast outputs can be edited in any way respecting 
those lines and still be used as input file for xblast.
In particular, several blast outputs can be merged, provided they
used the same query (e.g. blastx and blastn, tblastn and blastp).     

.SH DEFAULT
.LP
.TP
.B   xblast  <blast_output>  <query_filename> 
will generate a query masked with 'x',
.TP
.B   xblast  <blast_output>  <query_filename> N
will generate a query masked with 'N', etc ...
for using special masking characters like '!@#$%^&*- =><'
and others, quote them:
.TP
.B   xblast  <blast_output>  <query_filename> 'spcl_char'

for deleting instead of masking use empty double quotes:
.TP
.B   xblast  <blast_output>  <query_filename> ''
.TP
.B   | xblast  +  <query_filename>  <X_letter>
acts as a filter, and can be used directly after a blast search such in:
.TP
.B  blastx nrdb query S=60 | xblast  +  query_filename  > masked_query
 + is used as a place holder for the non-permanent blast output.
This option is useful within "filtering scripts".

.SH OPTIONS
.LP
.TP
.B   xblast -d  <blast_output>  <query_filename>
will generate a fasta-format database of the matched segments.
.TP
.B   xblast -r  <blast_output>  <query_filename>
will generate a "reversed" output, e.g. masking the non-matched segments.
.TP
\" When ever possible, pointers to related commands should be given.
.SH SEE ALSO
.LP
.I
XNU : a low-complexity sequence filter
DFBLAST : a fasta sequence annotator
the blastp/blastn/blastx/tblastn program suite

.SH REFERENCES
.nf
Jean Michel Claverie & David J. States (1993)
Information Enhancement Methods for Large Scale Sequence Analysis,
Computers and Chemistry 17: 191-201.

Jean Michel Claverie (1993)
Large Scale Sequence Analysis, Chapter 36 in
"Automated DNA sequencing and analysis techniques"
(J.C. Venter, ed), Academic Press

\" This section can be real useful.
.SH BUGS
.LP
Please report grateful comments, bugs or improvement suggestions to 
the author.

.SH AUTHOR
.nf
Jean-Michel Claverie                    jmc@ncbi.nlm.nih.gov
National Library of Medicine              (301) 496-2475
8600 Rockville Pike,  38A 8S806
Bethesda, MD  20894
.fi