Design of proteins by parallel tempering in the sequence space

Abstract

Design of new proteins is often formulated as an optimization task. An amino acid sequence is characterized by an energy, and this energy is sampled and minimized. Here, we use a parallel tempering algorithm to accelerate this task. A series of 100- or 200-residue proteins was designed using a modified Evolutionary Scale Modeling design module to maximize the confidence in structure prediction and globularity and minimize the surface hydrophobic residues. We show that parallel tempering is a viable alternative to Monte Carlo sampling and simulated annealing or related energy-based protein design methods, especially in the situation where a continuous flow of designed sequences is desired.