Nigel asks what blaze is - well here's the info sheet.
Dan Jacobson
----------------------------------------------------------------------------
GenBank now offers a preliminary version of BLAZE, a massively parallel
sequence similarity search program from IntelliGenetics and MasPar Computer
Corporation.
BLAZE - full version
--------------------
The full version of BLAZE - currently under development - searches DNA and
protein databases for sequence similarity and reports the alignment score,
percentage of match, and statistical significance of each database sequence.
Queries can include stop codons and IUPAC codes. BLAZE also can report the
optimal Smith-Waterman local alignment of your query with a database sequence
by showing the number of matches, the number of gaps, and the number of
mismatches. BLAZE can also display the annotations.
Standard supported databases are GenBank, Swiss-Prot, GENESEQ, VectorBank,
PIR and HIV. User-developed databases are also supported in the FastA
(Pearson) or IntelliGenetics formats.
BLAZE uses the full dynamic programming algorithm of Smith and Waterman for
maximum sensitivity. Every character in the query is compared with every
character in the database. At every position, BLAZE considers every possible
insertion or deletion gap of any length. Approximate methods employing
word matching are not used.
BLAZE - preliminary version
---------------------------
The preliminary version now being offered contains important limitations to
the functionality described above. Here are the current limitations:
1) The dynamic programming algorithm currently uses a single penalty for
insertion or extension of gaps. The full version will allow both a gap
and a gap-extension penalty.
2) The local alignments are not displayed.
3) The only supported database is Swiss-Prot.
4) Statistical significance is not printed.
BLAZE - anticipated schedule
----------------------------
The full version of BLAZE is expected to be released in the 4th week of July.
BLAZE - performance
-------------------
Executing on a MasPar MP-1 massively parallel computer, BLAZE can search
sequence databases 1000 times faster than on a workstation, comparing tens of
millions of residues per second. The 4,096 processor MP1104 system running
BLAZE searches Swiss-Prot 21 in 15 seconds with a query of 100 amino acids.
Access to BLAZE
---------------
You can access the GenBank BLAZE Server through a number of different
networks, including Internet, BITNET, EARN, NETNORTH and JANET.
The BLAZE program allows you to send a specially formatted mail message
containing the protein query sequence to the BLAZE Server at GenBank.
A BLAZE sequence similarity search is then performed using the full Smith
Waterman dynamic programming algorithm.
If you use BLAZE as a research tool, we ask that the IntelliGenetics/MasPar
BLAZE software be acknowledged in any resulting publication or public
presentation.
If you find errors, inconsistencies or other problems using the BLAZE software
we ask that you notify IntelliGenetics or GenBank via electronic mail.
To access the program, send an electronic mail message containing the formatted
query sequence (as described below) to the following Internet address:
BLAZE at GENBANK.BIO.NET
If you are not on Internet, you may need to change the format of the address.
Consult your systems manager to determine the correct address.
Obtaining Help
--------------
To receive instructions on using the BLAZE program, send a mail message
containing the word "Help" on a single line of the mail message to the address:
BLAZE at GENBANK.BIO.NET
Leave the Subject line in the mail header blank. The help text will be
updated when new information is available for BLAZE searches (such as new
databases on-line).
To request additional help concerning BLAZE, send an electronic mail message
to the address:
BLAZE-REQUEST at GENBANK.BIO.NET
Items sent to this address will be answered by a BLAZE support person.
