Algorithmic Research Problems in Molecular
Bioinformatics
Thomas Lengauer
Gesellshaft fur Mathematik und Datenverarbeitung mbH Schlosch Birlinghoven
D-53 757 Sankt Augustin, Germany
Department of Computer Science, University of Bonn, Germany
Abstract
The enormous amount of data that is provided by the genome sequencing projects around the world poses a great challenge to science, namely, to decode and interpret these data. There are different aspects to this problem, such as
- resolving the three-dimensional conformation of proteins and nucleic acids known by their sequence,
- uncovering the mechansims underlying the biological function of molecules such as proteins and nucleic acids,
- gaining insights into the role of non-coding stretches of DNA.
Advances in computing technology encourage us, to approach some of these questions with the aid of the computer. Due to the dramatic increase in computing power and, in particular, due to the newly gained graphics capabilities, we can visualize structure as well as dynamics of biopolymers on the computer screen. What is still largely missing is a set of reliable models and algorithmic methods for inferring molecular conformation on the basis of sequence knowledge and for reliably predicting the chemical interactions among macromolecules and between macromolecules and smaller ligands.
In this talk, we want to address a number of problems, whose solution requires algorithmic research as an essential ingredient. As a result, cooperation with computer scientists may be a promising approach to solving these problems. In response to this challenge, the German science ministry has started a research initiative entitled "molecular bioinformatics" that supports interdisciplinary research in this area incorporating molecular biologists and computer scientists over a period of four years.
We plan to give a general overview over a whole set of problem areas in the field of molecular bioinformatics and then focus on problems in the areas of homology-based modelling and docking.