Chapter 13 THREADER: protein sequence threading by double dynamic programming

Abstract

This chapter presents “THREADER”—protein sequence threading by double dynamic programming. The prediction of protein tertiary structure from sequence may be expressed symbolically by expressing the folding process as a mathematical function, C = F(S). The folding process is a function that takes an amino acid sequence and computes from it, a sequence of main chain torsion angles. The choice of representation of the folded chain conformation in torsion space is arbitrary and the problem can be expressed in terms of relative orthogonal 3D coordinates or with some indeterminacy in chirality, interatomic distances. The protein folding problem can thus be considered a search for the folding function F. However, it is probable that no simple representation of the folding function exists, and that even if the function exists in any form whatsoever, the only device capable of performing the required function evaluation is the protein chain itself. The simplest way to arrange for a protein sequence to code for its own native 3D structure is to arrange for the native structure to be the global minimum of the protein chain's free energy. The folding process is transformed into an energy function minimization process, where the energy function can take as input the protein sequence vector S, and the vector of torsion angles C .