- News
- Scientific projects
- Grid computing
- Decrypthon program
- Role-players
- Partners
- Thierry Lhermitte godfather
January 2007 - An interview with Alessandra Carbone
Mathematician, Alessandra Carbone is a computer science teacher and researcher at the “Université Pierre et Marie Curie – Paris VI”. Her research combines mathematics, algorithmic, and molecular biology. She is developing tools to identify protein-protein, protein-ligand and protein-DNA interaction sites.
Can you describe the highlights of your work?
Alessandra Carbone: The Inserm team I manage at the “Faculté de Médecine de Paris VI” studies the relationships between mathematics, algorithmic and molecular biology. We work on different problems related to the functioning and the evolution of biological systems. We are particularly interested in the description of genetic and biochemical networks in terms of systems. The approach is based on the analysis of experimental data and theoretical modelisation. It uses mathematical tools, notably statistics and combinatorial analysis, and algorithms to study the basic principles of cellular functioning, with genomic data.
A.C.: To identify these sites, two methods exist. “Molecular Docking” consists of using two proteins. One is immobile while the other revolves around the former to identify the positions of interaction. The second approach, called “Evolutionary trace”, flattens the protein. In other words, contrary to Molecular Docking which analyses proteins in 3D, this method is only based on the amino-acid sequence (note: protein components). Using the knowledge already acquired about interaction sites, for a given protein, researchers determine the areas in its amino-acid sequence with the highest probability of belonging to an interaction site. Then, as in the first method, we study what happens with the protein in 3D. Thanks to these two approaches, it will be possible to identify the protein interaction sites and especially false negatives which mislead researchers.
There must be a countless number of possibilities?
A.C.: For molecular docking, there are approximately 30 000 different positions to test. With a 200 second computing time for each position, it takes several days to analyse only one protein. So what can be said about the thousands of proteins around us? By carrying out a first selection of relevant sites by using the evolutionary trace method, we decrease dramatically this time, but there are still 300 potential positions to test.
Hence the interest of a tool such as Decrypthon?
A.C.: Indeed, the computation power offered by Decrypthon theoretically enables us to speed up protein analysis. We should continue with the idea of shared computation. Finally, no matter what computing tool is chosen, we still have a lot of work ahead of us to develop the program which will enable us to demonstrate the reliability of the two methods, molecular docking and evolutionary trace, on 85 protein complexes.
