A Multisequence Monte Carlo Method in a Simple Protein Model

In this thesis an algorithm for simulating the equilibrium behavior of a large number of protein sequences at fixed temperature in a single run is investigated. The method, which works by allowing Monte Carlo updates of the protein sequence, is implemented and tested for a simple reduced-representation continuous protein model with three different types of amino acids. The effectiveness of the model is related to that of a naive fixed-sequence Monte Carlo algorithm, by comparing thermalization times and statistical errors. The new algorithm is found to have considerably shorter thermalization times, and statistical errors of similar or lower magnitude for most sequences investigated. Finally, the method is used to investigate the sequence-structure relationship for short model proteins with eight amino acids. Sequences folding to certain structures were mapped, and the number of sequences folding to each were determined, as well as the number of mutational transitions between them