Efficient modeling of supra-molecular systems using GPGPU
Massively parallel molecular simulations can be used to understand the flexibility of huge macromolecular systems . Very long time-scale molecular simulations  (on the order of milliseconds) are often required to accurately model several key biological phenomena. Modern Graphics Processing Units (GPUs, the components of a graphics card that executes instructions for an informatics program) can be used for general purpose calculations [3,4,5]. This tremendous potential encouraged computer scientists to design new algorithms for massively parallel execution on the GPU.
In this context, I have worked on high performance (parallel) and hybrid (GPGPU) computing, benchmarking and using codes such as NAMD and GROMACS on huge hybrid (CPUs/GPUs) clusters hosted by the CCRT.
Efficient visualization of supra-molecular systems using GPU capabilities
Using such GPU capabilities, several recent studies have redesigned traditional algorithms to exploit new graphics card capabilities, hence radically improving display performance [6,7,8]. When I was a PhD student, I have designed, with supervisors B. Maigret and B. Levy, an algorithm to display molecular surface changes in real time . Using GPU ray-casting, this program improves both quality and efficiency of molecular surface rendering as shown above.
I have also developed a new representation called HyperBalls  during my post-doc which can be used as an alternative for representation of Coarse Grain (CG) models, using smooth links that provide more consistency. The representation is based on hyperboloids, and is dependent upon atomic radii, inter-atomic distances and user-operated sizing parameters. For example, the use of HyperBalls is well adapted to display lipid bilayers and especially to depict their crowded environment as well as the compactness of the membrane (see movie section on HyperBalls website). The HyperBalls representation is also suited for the visualization of spring networks: the thickness of the bonds can be modified as a function of the distance between atoms, hence depicting the spring behavior and handling the visual creation of bonds dynamically This characteristics makes it very useful for monitoring the dynamics of hydrogen bond breaking and formation (including water-mediated hydrogen bonds), wherein one can observe the continuous evolution of bond formation by monitoring the thickness of the bonds, which increase in size as the bond forms and vice-versa.
More information on HyperBalls website.
Modeling Protein-Protein and protein-Ligand Interactions
I have used bioinformatics programs to model the interactions between Erbin PDZ and Smad3 MH2 domains, in collaboration with biologists of the Paoli-Calmettes Institute: a cancer research institute in Marseille . I have used Molecular Dynamics (MD) simulations to refine docking results from researcher Dave Ritchie during the CAPRI experiment. We obtained particularly good results for the blind Target 34: protein-RNA interaction (see P61 results).
I also participated into the VSM-G virtual screening project, the aim of which is to filter large databases of molecules in order to identify possible hit compounds .
1. Freddolino PL, Arkhipov AS, Larson SB, McPherson A, Schulten K (2006) Molecular dynamics simulations of the complete satellite tobacco mosaic virus. Structure 14: 437-449.
2. Shaw DE, Deneroff MM, Dror RO, Kuskin JS, Larson RH, et al. (2008) Anton, a special-purpose machine for molecular dynamics simualtion. Communications of the ACM 51: 91-97.
3. Stone JE, Hardy DJ, Ufimtsev IS, Schulten K (2010) GPU-accelerated molecular modeling coming of age. J Mol Graph Model 29: 116-125.
4. Dematté L, Prandi D (2010) GPU computing for systems biology. Briefings in Bioinformatics 11: 323-333.
5. Stone JE, Phillips JC, Freddolino PL, Hardy DJ, Trabuco LG, et al. (2007) Accelerating molecular modeling applications with graphics processors. J Comput Chem 28: 2618-2640.
6. Krone M, Bidmon K, Ertl T (2008) GPU-based Visualisation of Protein Secondary Structure. In: Lim IS, Tang W, editors. Theory and Practice of Computer Graphics. pp. 1-8.
7. Daae Lampe O, Viola I, Reuter N, Hauser H (2007) Two-level approach to efficient visualization of protein dynamics. IEEE Trans Vis Comput Graph 13: 1616-1623.
8. Sigg C, Weyrich T, Botsch M, Gross M (2006) GPU-Based Ray-Casting of Quadratic Surfaces. In: Botsch M, Chen B, editors. Eurographics Symposium on Point-Based Graphics.
9. Chavent M, Levy B, Maigret B (2008) MetaMol: high-quality visualization of molecular skin surface. J Mol Graph Model 27: 209-216.
10. Chavent M, Vanel A, Tek A, Levy B, Raffin B, et al. GPU-accelerated dynamic of molecular interactions and representations as HyperBalls, a unified algorithm for balls, sticks and hyperboloids. Accepted in Journal of Computational Chemistry
11. Déliot N, Chavent M, Nourry C, Lécine P, Arnaud A, Hermant A, Maigret B, Borg JP (2008) New Insight into the interaction between erbin and smad3: a non-classical binding interface for the erbin PDZ domain. BBRC 378(3):360-365.
12. Beautrait A, Leroux V, Chavent M, Ghemtio L, Devignes MD, Smail-Tabbone M, Cai W, Shao X, Moreau G, Bladon P, Yao J, Maigret B. (2008) Multiple-step virtual screening using VSM-G: Overview and validation of fast geometrical matching enrichment. J Mol Mod, 14 (5): 393-401