Protein-protein docking algorithms consists in two successive steps: first a large number of putative conformations is generated by an exploration algorithm, then a scoring function is applied to rank them. As Critical Assessment of PRediction of Interactions (CAPRI) experiment shows, exploration procedures for protein-protein docking have really been improved in recent years but scoring functions are still a difficult issue.
Our work focus on the second part of this problem: finding a fast and reliable scoring function. This function has to take into account physico-chemical properties of the residues and geometric complementarity of the two protein partners. Whereas existing scoring functions considers these criteria separately, we decided to use Voronoï tessellations which have been shown to be an efficient way of describing protein structures. Voronoï construction allows the description of both protein geometry and residue properties.
Jérôme Azé, LRI, Orsay, France.
Alexandre Bonvin, Bijvoet Center, Utrecht, The Nederlands.
Frédéric Cazals, INRIA, Sophia Antipolis, France.
Marie-Hélène Mucchielli-Giorgi, CGM, Gif-sur-Yvette, France.
The CGAL Project (Computational Geometry Algorithms Library).
A new docking scoring function based on interface geometry and physico-chemical residue properties
CECAM workshop on Flexible Macromolecular Docking
Une nouvelle fonction de score pour l’amarrage protéine-protéine fondée sur les diagrammes de Voronoï
Julie Bernauer, Jérôme Azé, Joël Janin, Anne Poupon
A docking analysis of the statistical physics of protein protein recognition
Julie Bernauer, Anne Poupon, Jérôme Azé, Joël Janin
Physical Biology, 2005, 2, S17-S23
