Section: New Results

Combination of Evolutionary and Semantic Web Techniques for Protein Design

Participants : David Simoncini, Andrea Tettamanzi.

Proteins are fundamental components of all living cells and are among the most studied biological molecules. They are involved in numerous diseases and being able to determine their 3D structures and interactions is essential to understand the mechanisms of cell functions. De novo computational protein design refers to the problem of finding a sequence of amino acids corresponding to a protein with the desired three-dimensional structure, or the desired biological function. It is a longstanding goal in computational structural biology and only a few examples of successful de novo computational protein designs can be found in the literature. Computational protein design has many industrial applications, such as biofuels, drug synthesis and food processing (through computational design of enzymes) or targetted drug delivery systems (through bio-nanotechnologies).

In this context, our research focuses on knowledge extraction from protein structure databases for the development of new computational protein design frameworks. Whereas most of the current methods ignore available structural information, our algorithm takes into account known profitable interactions between amino acids and uses this information to guide the energy minimization process and propose more realistic sequences of proteins.