++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+ +
+ BBBBBB OOOOO SSSSSS CCCCCC AAA N N +
+ B B @ O O S C A A NN N +
+ B B i O O S C A A N N N +
+ BBBBBBB ii O O SSSSS C AAAAAAA N N N +
+ B B ii O O S C A A N N N +
+ B B ii O O S C A A N NN +
+ BBBBBB iii OOOOO SSSSSS CCCCCC A A N N +
+ +
+ _ _ ___ +
+ | | |_| _ ___________ _____ _____ __/ / +
+ | |__ _ ____ | |__ / _____/ // _ // _ // / +
+ | || || . || | /__ // / // __// ___// / / +
+ |__|_||_||__ ||__|_|/_____// __//____//____//____/ +
+ |_____| /_/ "Computer!" +
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
UNIVERSITY OF NORTH CAROLINA AT CHAPEL HILL
Department of Computer Science
CB #3175 Sitterson Hall
Chapel Hill, NC 27599-3175
Tel: (919) 962-1744
Fax: (919) 962-1799
o AUTOMATIC E-mail Server
-- bioscan at cs.unc.edu
o ANONYMOUS FTP Server for Client Software
-- Host: ftp.cs.unc.edu (presently 152.2.128.159)
-- Path: /pub/bioscan
o Other Information (Human interface)
-- bioscan-info at cs.unc.edu
============================================================================
BioSCAN: A Scalable Multiprocessor System for Biopolymer Similarity Analysis
============================================================================
o DESCRIPTION
BioSCAN (Biological Sequence Comparative Analysis Node) is a massively
parallel computer system designed for biological sequence similarity
analysis. This system performs rapid, rigorous, searches of DNA, RNA,
and protein sequence databases and is compatible with popular software
packages such as BLAST and FASTA. BioSCAN uses a rigorous implementation
of linear similarity methods used in the BLAST systems. It is also well
suited to multiscore applications using multiple PAM and BLOSUM scoring
matrices.
o OVERVIEW
o Speed is independent of query sequence length.
o System performance is scalable. Maximum query sequence length
is determined by the size of the system. (Current VMEbus based
design allows up to 12,992 characters in query sequence)
o Rigorous implementation of linear similarity method as used
in the BLAST system; well suited to multiscore search application
using multiple PAM and BLOSUM scoring matrices.
o CONCEPT
o Combination of hardware and software carefully chosen to
provide high performance without sacrificing flexibility.
o Simple linear similarity algorithm implemented in hardware
reports significant alignments in the database at high speed.
o Core algorithm is implemented in a full-custom VLSI chip.
o Layered software allows changes and further development of
system components without impacting the applications".
o FEATURES
o Suitable for DNA, RNA and Protein sequences.
o Can perform simultaneous searches with multiple LOG-ODDs matrices
such as PAM and BLOSUM.
o Can perform simultaneous searches of multiple query sequences.
o Software architecture supports stand-alone, Client-Server ,
and E-mail access to the system.
o An Application Programming Interface (API) provides easy interface
for third-party application software development.
o PERFORMANCE
o Each VLSI chip scans a maximum of 2 million elements (DNA and RNA
bases or amino acid residues) per second.
o System scans GenBank 77 database (size 139 million bases) in 71
seconds and SWISS-PROT 25 database (size 10 million residues)
in 5 seconds.
o GETTING STARTED
o To receive the current set of instructions on using the BioSCAN
e-mail server, send an e-mail to the address "bioscan at cs.unc.edu".
Put the word "HELP" on a line by itself in the body of the message.
o A tutorial can be obtained via anonymous FTP from "ftp.cs.unc.edu".
The directory is /pub/bioscan and the file is called "tutor.txt".
o Please direct problems with the server and other inquries to
"bioscan-info at cs.unc.edu".
* ACKNOWLEDGEMENTS
This research was supported in part by:
o NSF, under Grant No. MIP-9024585
o MCNC, under Design Initiative Research Program
==============================================================================
Prof. Raj K. Singh
Dept. of Computer Science
UNC Chapel Hill
CB # 3175 Sitterson Hall
Chapel Hill, NC 27599-3175
Off: (919) 962-1744l
Fax: (919) 962-1799
E-mail: singh at cs.unc.edu
