TITLE:

AUTOMATED CHEMICAL STRUCTURE DETERMINATION: ON THE PROPERTIES OF THE MONOSACCHARIDE GC-EIMS SPECTRAL PROBLEM SPACE (No. 100)

DATE:

Friday, January 27th, 2006

TIME:

3:30 PM

LOCATION:

GMCS 214

SPEAKER:

Faramarz Valafar, Department of Computer Science, San Diego State University

ABSTRACT:

Chemical Structure Determination (CSD) is an essential part of many
projects in Chemistry, Biology, and Medicine. Its applications are used
in a wide range of industries including law enforcement in CSI-type
laboratories for post-crime analysis. In recent years, CSD has been
propelled into prominence due to homeland security’s needs in prevention (as opposed to post-crime analysis) efforts.

Furthermore, CSD is a complex and time consuming task that requires a high level of specialization. What makes the task further difficult is
its repetitive and tedious nature, making the process prone to human
error. In service-oriented analytical chemistry laboratories, it is
common to have highly specialized staff in place for this highly
repetitive and manual process. Additionally, because the majority of the molecules that undergo structure determination are already known
molecules, the efforts in this area are highly duplicated in
laboratories all over the world, leading to a great deal of
unnecessarily wasted expert man-hours. Moreover, applications such as scanning luggage at airports for explosives have further raised the need for “real-time” or “online” CSD. To achieve this, Automated CSD (ACSD) was proposed. ACSD addresses many of the problems faced by CSD. It promises to eliminate the majority of human error from the process. It also promises to make the process fast and feasible for applications such as explosive scanners at airports. While ACSD has been successfully implemented for some (small set of) target chemical compounds, we still lack such systems for structural identification of a wide range of molecules.

In this project, we discuss one such system for identification of all
monosaccharides that exist in nature from their GC-EIMS spectra. In the process, essentially, we will also discuss the mathematical properties of the GC-EIMS problem space for all monosaccharides. We will discuss some possible pattern recognition solutions to the problem of ACSD of monosaccharides, and report our current success rate in doing so. In the talk, we will also briefly discuss a web-based implementation of this system, which we have called “I-Chemist” (Intelligent-CHEMIcal STructural database) (http://zal.sdsu.edu/i-chemist/).

HOST:

Paul Paolini

DOWNLOAD: