Entropy functions and rare events in disordered systems by transfer matrix calculations and Monte Carlo sampling

This thesis treats problems from bioinformatics with Monte Carlo methods from statistical physics. Methods to compare molecular sequences (sequence alignment) make use of statistical tests to assess the significance of observed similarities. Distributions of optimal alignment scores over random sequences are integral parts of such tests. I extended an existing work where large deviations from the asymptotically predicted Gumbel distribution were found to further biologically relevant scoring and protein models. In most cases, deviations could be described by an heuristically modified Gumbel distribution. In a similar way, I studied the minimum free-energy distribution of RNA secondary structures over ensembles of random sequences. Furthermore I investigated the dynamics of different Monte Carlo algorithms in complex energy landscapes and the connection to static properties.

Level thesis:degree_level

thesis.doctoral

Grantor dc:publisher

Universität Oldenburg

Year

2009

Author dc:creator

• Wolfsheimer, Stefan

• [Keine Schlagwörter von Autor/in vergeben.]

Repository record source_url

http://oops.uni-oldenburg.de/838

OAI identifier oai:identifier

oai:oops.uni-oldenburg.de:838