EU-Projekt: Emergent organisation in complex biomolecular systems, EMBIO - Specific Targeted Research Project,
NEST – EC Contract No. 012835
Description of the project:
Complexity and self-organisation are critical yet poorly understood phenomena. This project aims to develop and apply mathematical and computational approaches that will identify principles governing the emergent organisation of self-organising biomolecular systems. Computational methods for characterising the dynamics of these intrinsically complex processes will be developed and applied to protein folding and molecular self-assembly. The methodologies will focus on the complexitiy of the system’s dynamics thus advancing fundamental knowledge concerning the role of complexity in biological systems.
In this project, we studied properties of energy landscapes of proteins and RNAs. We investigated how to (efficiently) measure features of the landscape such as the number of local minima, the size distribution of the basins of attraction, the barrier tree, etc. A number of techniques for this purpose have already been developed for the special case of RNA secondary structure [46, 47]. Most of them depend crucially on the possibility to enumerate the near-ground state part of the energy landscape. The structure of the barrier tree itself can be used to extract characteristic quantifiers. A simple approach  will be extended to more sophisticated tree descriptors. We propose here to develop similar tools that can be applied to lattice models of protein folding. In [38,39,40], we have designed a global optimization technique for HP-kind lattice models on the cubic and face-centred cubic lattices based on constraint programming approaches. The HP-model is a simplified protein model introduced by Ken Dill. It distinguishes between polar (P) and hydrophobic (H) amino acids, and searches for a conformation with a maximal packing of the hydrophobic (H) amino-acids. Using the constraint-based approach, we can successfully fold sequences up to length 300, thus greatly improving on the results of competing groups.
We want to use and extend this method to generate also the near-ground state part. One aim is to enlarge the sequence length which can be accessed using this method. But still we have to cope with an exponentially increasing number of conformations. For this purpose, we want to combine stochastic search methods with constraint-based approaches. This "constraint-based sampling" allows us a more directed (targeted) investigation of the landscape, thus improving both quality and efficiency of the sampling.
Additional information: http://www-embio.ch.cam.ac.uk
contact person: Martin Mann
Start of project: 01.03.2005
End of project: 31.08.2008
Prof. Dr. Rolf Backofen
Prof. Dr. Rolf Backofen
Phone: +49 (0) 761-203-7461
Fax: +49 (0) 761-203-7462
Actual Research Report