Protein Folding Simulation in CCP

A protein is a list of linked units called aminoacids. There are 20 different kinds of aminoacids and the typical length of a protein is 100–500 units. The Protein Structure Prediction Problem (PSP) is the problem of predicting the 3D native conformation of a protein, when its aminoacid sequence is known. The process for reaching this state is called the protein folding. This work deals with ab-initio prediction, based on Anfinsen thermodynamic hypothesis [1] which states that the conformation adopted by a protein (also known as the native conformation) is the one with minimum free energy. We can identify two main problems: the first is to choose a representation of the protein and an energy function, which must be at minimum for native-like conformations. The second is, given the representation and the energy function, to find the 3D conformation that minimizes the function. Due to intrinsic computational limits, no general solution to the latter problem is currently available. In particular, simulation-based techniques that take into account all atoms constituting the aminoacids (and the solvent) and simulate the folding process approximating atom interactions, run extremely slow due to the huge number of calculations.