[an error occurred while processing this directive] [an error occurred while processing this directive]

University of Saskatchewan, Department of Computer Science

2001-2002 Seminar Series

Bioinformatics and Biocomputing - Intersections in the Structure Prediction Domain

Kay Wiese
Assistant Professor of IT
Technical University of British Columbia

DEPARTMENT SEMINAR
DATE: Monday, December 3, 2001
TIME: 3:30pm
PLACE: Anthropology 132
*** Everyone is welcome ***

Abstract

In this talk I will begin with a brief introduction to the fields of Bioinformatics and Biocomputing and clarify the distinction between the two fields. Bioinformatics involves the development and application of advanced and innovative computational methods in order to address problems in molecular biology. Many of these problems are combinatorial or constraint optimization problems. Two problems that fall into this category are RNA and Protein Structure Prediction. Biocomputing involves the design and application of algorithms that are gleaned from natural processes such as evolution (Genetic Algorithms, Evolution Strategies), the functioning of a brain (Neural Networks) or using real DNA molecules to perform a computation (DNA computers).

I will introduce the problems of RNA and Protein Structure Prediction. These problems are of relevance since the structure of these molecules largely determines their function in the cell. Traditionally, structure prediction had to be performed in a wet lab using X-crystallography and NMR methodologies. This process is costly and time consuming. I will explain how the structure prediction problem can be decomposed so that it becomes an energy minimization problem and how we are using a genetic algorithm to solve it.

During the talk, I will also discuss various modifications to the genetic algorithm, with which we had success in the TSP domain and how we are now applying these modifications in the structure prediction domain. While the driving force behind an optimization algorithm for structure prediction is energy minimization, the real measure of success is how closely the predicted structure resembles the real structure. Here, visualization of predicted structures of bio-molecules can help to establish how well an algorithm can predict a known structure. Visualization is also important for biologists to conduct their own investigations using the predicted structures. I will briefly discuss one of my visualization projects in a 3D immersive environment (Cave) and time permitting, I will give a live demo of a protein visualization tool.

About the speaker

Kay C. Wiese has studied Mathematics and Computer Science at the University of the Saarland in Saarbruecken, Germany and the University of Regina, Canada where he received his PhD in 1999. After teaching Computer Science at UBC for one year he has joined the Technical University of British Columbia in June 2000 and currently holds the position of Assistant Professor in Information Technology. [an error occurred while processing this directive]