.\"manual page for xsact
.\"don't leave home without it
.\"
.TH XSACT 1 "FEBRUARY 2006"
.SH NAME
xsact \- cluster EST sequences
.SH SYNOPSIS
.B xsact -k
.I kval
.B -n 
.I threshold
.B [options] [file..]
.SH DESCRIPTION
.B xsact
Clusters ESTs in a FASTA-formatted file by building a suffix
identifying exactly matching blocs, and measuring the maximal
colinear set of matches between pairs of sequences.
.SH OPTIONS
.IP -k
Specify the word size
.IP -n 
Specify threshold for clustering
.IP -p
Prefix length.  Increase this to reduce memory consumption.
.IP -u
Upper case input only, lower case characters are treated as 'N's.
.IP -s
Add singleton clusters to output.
.IP -i
Don't ignore matches in simple (low complexity) repeats.
.PP Output options
.IP -o
Specify output file.
.IP -L
Each cluster is output as one line, containing its sequence labels.
.IP -N
Output a set of newick-formatted trees.
.IP -U
Output full sequences, each cluster separated by a comment line
(similar to UniGene).
.IP -P
Output all sequence pairs.
.PP Parallelism options
.IP -m 
Specify a prefix for which to generate intermediate files.
.IP -c
Construct clustering from intermediate files with the specified prefix
length. 
.SH EXAMPLES

Typical usage:

      xsact -k 20 -n 70 -s -x -p 2 input.seq -L -o input.clusters

Good values for k and n are in the ranges 18..30 and 50..90,
respectively.  For anything beyond trivial data sets, use 
.B -x
and set
.B -p
between 1 and 3.

Here's a script that clusters the file using four parallel processes:

      for a in A C G T; do
         xsact -k 20 -p 2 -m $a input.seq &
      done
      wait
      xsact -k 20 -n 70 -c 1 -s -L input.seq -o input.clusters

.SH BUGS
Large intermediate files are generated.  Keep an eye on free disk space!
.SH AUTHOR
Ketil Malde <ketil@ii.uib.no>