[Apologies for multiple postings]


Brian Chen's talk is scheduled tomorrow.
Please see details below.

**************************************************
*
*
*    GRAND Seminar
*
*    http://cs.gmu.edu/~robotics/Main/GrandSeminar
*
*
**************************************************


*Title*
An algorithm for discovering steric influences on
protein-ligand binding specificity

*Time/Venue*
CS conference room, ENGR 4201
Noon, October 11, Tue.

*Speaker*

Brian Chen
http://www.cse.lehigh.edu/~chen/
Assistant Professor
Dept. of Computer Science and Engineering
Lehigh University

*Host*

Amarda Shehu
http://www.cs.gmu.edu/~ashehu

*Abstract*

Living things are composed of interacting and nested systems
that exhibit symptoms of health and disease at all levels.
At the most fundamental level, understanding the correct
behavior of healthy systems, and resolving the breakdowns
that occur in disease, requires a precise understanding of
how proteins function and malfunction. Building such an
understanding is much like the examination of a complex
machine: the analysis of protein shape can yield many
insights, especially at functional sites, where biochemical
activity actually occurs.

Many algorithms have been designed to gather structural
observations, especially about the functional site, that
point to functional hypotheses that can be ultimately tested
on the bench. In every case, these methods employ a "guilt
by association" approach by assigning biological function
based on functional site similarity or a resemblance to
established norms. This talk asserts that more informative
approaches are possible and presents VASP, a new method that
dissects functional sites to isolate individual components
that play influential functional roles. VASP enables these
new capabilities by exploiting a novel connection to
concepts from computer graphics and computer aided design.

The results presented in this talk describe a case study
where VASP was applied to the analysis of the major serine
proteases. VASP isolated individual amino acids and regions
of functional sites that enable the serine proteases to
preferentially bind specific molecules. None of these
observations are possible with existing unsupervised
methods. These results point to applications in molecular
bioengineering, where the identification of individual
functional components is essential for the rational
engineering of active biomolecules, and to applications in
structure-based drug design, where important functional
components could be exploited to mitigate side effects or
evade drug resistance.



-- 
*Jyh-Ming Lien*
Assistant Professor, George Mason University
+1-703-993-9546
http://cs.gmu.edu/~jmlien