Halfwise and Blastwise

The use of genewise in large scale analysis is beyond most people's CPU abilities. To counter this I have written two scripts which allow people to use genewise more sensibly.

• Halfwise - a Perl script that compares a DNA sequence to Pfam sensibly, using BLAST to speed up the process.
• Blastwise - a Perl script that compares a DNA sequence to a protein database, using BLASTX and then calls genewise on a carefully selected set of proteins

To run halfwise you will need

• The Wise2 package, compiled to provide the genewisedb executable at least
• One of the blastx type programs, either blast 1 series, blast 2 series from ncbi or wublast from warren gish (bioperl automatically detects the different flavours of blast and adjusts).
• The bioperl distribution, preferably the 0.05 series
• The halfwise protein database, found at ftp://ftp.sanger.ac.uk/pub/birney/wise2/halfwise
• The halfwise Pfam database, at the same ftp site
• The HMMER package, version 2.1 series

The halfwise database is made from the Pfam FULL alignments, made non redundant to 75being quite a small database.

To install halfwise you need to

• place the halfwise protein database in the directory pointed by BLASTDB and either pressdb or setdb depending on which version of blast you are going to use. (If you don't have the BLASTDB directory set up, make a directory called blastdb and set the environment variable BLASTDB to point to that)
• install bioperl, best by following the instructions in the README
• install Wise2
• install the latest HMMER
• place the HMM library in BLASTDB and run hmmindex (from the HMMER package) on it
• edit the information at the top of the halfwise.pl to point to the correct executables, if need be

To run halfwise go

halfwise dna.seq > dna.seq.hlf

halfwise by itself gives you help about it.

To run blastwise you will need

• a blastable protein database
• one of the blastx type programs, as above
• The Wise2 package, having made the perl port. This is done by going ``make perl'' in the root directory
• Bioperl version 0.05 or above
• a way of fetching fasta formatted sequences from teh protein database, eg SRS

Install bioperl and blast as before, install the Wise2 perl port. Edit the blastwise.pl script, making sure you change protein database and the GETZ line lower down to represent the way of getting sequences.

To run blastwise go

blastwise.pl dna.seq > dna.seq.blw

The blastwise script is designed to be adjusted to fit your site. There are a number of us world wide concentrating on extending and improving blastwise. Please get in touch if you want to help.