New Techniques for Protein Simulations the Generalized-Ensemble Approach

The energy landscape of proteins and peptides is characterized by a multitude of local minima separated by high energy barriers. Low temperature simulations by canonical molecular dynamics or Monte Carlo will get trapped in configurations corresponding to one of these local minima. We show how this problem can be overcome by new simulation techniques such as generalized-ensemble algorithms. We demonstrate the effectiveness of this approach for the structure prediction of peptides and proteins, and study these molecules from a statistical physics point of view. As an example, we investigated the helix-coil transition through calculating partition function zeros and finite-size scaling analysis. Especially, we were able to determine a set of critical exponents which characterize the coil-helix transition. © 2001 IOS Press.