Item

Abstract

Proteins play a crucial role in nearly all biological processes ranging from cell division to its death. Despite the variety of the functions they perform, the chemistry of the protein world is rather simple, encoded by an alphabet of 20 amino acids. The order of amino acids in a protein sequence determines its structure, which, in turn, defines the function. For a protein chain of a typical length, the number of possible amino-acid sequences is essentially infinite; however, only a tiny subset of this space was tested by nature and its majority remains unexplored. Computational protein design methods allow to “visit” the protein universe “dark matter” and thus create new proteins of previously unseen structures and functions. In one of our projects, we attempt to design a “pi-helical coiled coil”, a novel protein fold composed of two naturally occurring structural motifs – rare and unstable pi-helices and ubiquitous and stable coiled coils. Here, we present new bioinformatics tools that we have developed in the course of this project, namely neural network-based methods for accurate prediction of pi-helices and coiled coils in protein sequences and new protein design protocols, involving molecular dynamics simulations, enabling the exhaustive exploration of the sequence space.