IUBio Biosequences .. Software .. Molbio soft .. Network News .. FTP

Announcing CarbNet, A Web Based Analytical Ssystem

Faramarz Valafar faramarz at MOND1.CCRC.UGA.EDU
Mon May 13 09:21:49 EST 1996


The Complex Carbohydrate Research Center (CCRC) of the 
University of Georgia is announcing an analytical 
glyco-chemistry service via the World Wide Web.  The software is 
called CarbNet (Carbohydrate Neural Networks), and is available 
at http://www.ccrc.uga.edu/ CarbNet is an analytical tool for 
identification and verification of carbohydrate spectra.  The 
system uses a neural network pattern matching engine to match 
spectra of unknown compounds submitted by the user with spectra 
of compounds from CCRCs databases.  Each search of CarbNet 
produces a ranked "hit list" of compounds in our library whose 
spectra most closely match the unknown spectrum, and a graphical 
presentation of both the submitted spectrum and the "best match" 
sample spectrum from CarbNet.  Each compound in the hit list is 
associated with a score that represents the closeness of the 
match.  The graphical presentation assists in the visual 
verification of the search result.  A capability is provided for 
"static" zooming into a certain region of the spectra.  

The CCRC currently offers analytical services for two types of 
complex carbohydrates.  These are: combined Gas 
chromatography-electron impact mass spectra (GC-EIMS) of 
partially methylated alditol acetates (PMAAs), and for 500-MHz 
NMR spectra of xyloglucan subunit oligosaccharides.  

The CarbNet software is currently a prototype.   We hope that, 
after using this prototype, your feedback will assist us in 
developing an improved version.  Some improvements such as 
"dynamic" zooming capability will be incorporated as soon as 
JAVA (or an alternative such as OLE) compatible web browsers 
become more comon.  CarbNet currently uses some extensions to 
HTML 2.0.  Therefor, it is best presented by Netscapes or 
Microsofts web browsers.  

If you have a spectrum of the above-mentioned types, we 
encourage you to submit it to CarbNet and let us know how well 
CarbNet performs in helping to identify your compound.  We 
especially encourage you to send us an e-mail message if you 
encounter a problem or receive an incorrect response from 
CarbNet.  You should be aware that these spectral libraries are 
not complete; we plan to add spectra to each of these libraries 
and to introduce libraries of other types of complex 
carbohydrates.  If either library does not produce a "match" for 
your submitted spectrum, we ask that you contact us by e-mail at 
the address below so that we can arrange to add your spectrum to 
our library.  Thereby increasing the value of the library to the 
scientific community.  

On the system, we have sample spectra for above mentioned 
libraries.  If you do not have a spectrum to submit, we 
encourage you to use these spectra to initiate a search, and 
observe the functionality of the system.  We welcome any 
comments that you might have.

For further inquiries or any problems with CarbNet, please 
contact faramarz at mond1.ccrc.uga.edu.


-- 
Faramarz Valafar
Complex Carbohydrate Research Center
University of Georgia
220 Riverbend Rd.
Athens, GA 30602
USA
Tel: (706) 542-4436
Fax: (706) 542-4412



More information about the Glycosci mailing list

Send comments to us at biosci-help [At] net.bio.net