The algorithmics of folding proteins on lattices

Abstract

It should be possible to predict the fold of a protein into its native conformation, once we are given the sequence of the constituent amino acids. This is known as the protein structure prediction problem and is sometimes referred to as the problem of deciphering the second half of the genetic code. While large proteins fold in nature in seconds, computational chemists and biologists have found that folding proteins to their minimum energy conformations is a challenging unsolved optimization problem. Computational complexity theory has been useful in explaining, at least partially, this (Levinthal's) paradox. The pedagogic cross-disciplinary survey by Ngo, Marks and Karplus (Computational Complexity, Protein Structure Prediction and the Levinthal Paradox, Birkhauser, Basel, 1994) provides an excellent starting point for non-biologists to take a plunge into the problem of folding proteins. Since then, there has been remarkable progress in the algorithmics of folding proteins on discrete lattice models, an account of which is presented herein.