A quantum walks assisted algorithm for peptide and protein folding prediction.

Proteins are considered as the working force of cells. Their functionality is determined by their spatial form. In 1973 Anfinsen proposed that the spatial form is determined by the sequence of amino acids in the protein backbone. Yet, the number of possible sequences as well as the possible configurations is very large, making the task of predicting the protein's spatial form very difficult. Many approaches have been proposed, both classical and hybrid quantum - classical ones. We propose a novel hybrid algorithm. In our approach we utilized quantum walks, a proven model for universal quantum computation. We considered a simplified version of the protein backbone to be the evolution space of the quantum walk. The dihedral angles φ and ψ are introduced as phase factors to the quantum walk evolution. We also utilized a cost function to describe the system, where the R - chain, describing the specific amino acid, corresponds to a discrete value, affecting the cost functions value. Our aim is to minimize the cost function value, by updating the dihedral angles for specific regions of the Ramachandran plot, using a Metropolis algorithm.