In 2005 the Blue Brain Project (http://bluebrain.epfl.ch), headed by Prof. Henry Markram, started out to introduce simulation-based research in neuroscience, the ability to research the fundamentals of neurobiology in silico. The Blue Brain Project provides a comprehensive facility to generate and validate biophysically accurate models of cortical brain tissue at the cellular level. The core of the facility is a tight iteration cycle between experiment, model building, in silico experimentation and automated validation that can be completed in a week's time on the CADMOS BlueGene/P (16k cores) supercomputer. The current iteration is based on more than 10 years of quantitative electrophysiological recording in the somatosensory neocortex of a young rat and presently consists of 10,000 morphologically detailed neurons with genetically constrained ion channel populations distributed across the 3D morphology. This facility now provides the first-of-a-kind opportunity to perform rapid experiments on this virtual tissue that would require complex technical apparatuses and years to carry out on biological tissue. Already many biological findings can be reproduced using virttual tissue and that even deeper insights can be obtained than possible on biological tissue. The facility can now also be used for hypothesis testing, where proposed alterations in disease can be integrated and the hypothesized disease state can be simulated.

To study the genome evolution of species, one can look at coevolving positions of genes. Coevolution is defined as "the modification of a biological object triggered by the change of a related object". It has been described in various biological systems and can be an essential process behind changes occurring at both morphological level, e.g. coevolution of female and male genital morphology and molecular level, e.g. ligand-receptor interactions.

We recently published a model, Coev, that accurately identifies the coevolving positions and estimates the associated co-evolving profile. The principal focus of this work is the application of Coev model on a large number of protein to better understand the intrinsic properties of coevolving positions in vertebrates.

The Vital-IT High Performance Computing Center supports a wide range of different life science applications on its heterogeneous HPC cluster. Several of the applications also use MPI to provide speedup execution time. The goal of the project is to benchmark a set of life-science applications (including adapting the code to BlueGene/P where necessary) in order to test how applications scale to a large number of processing cores and high-speed interconnect. This knowledge will be fed back to the Arc Lémanique scientific user community to find optimal execution environments for specific applications in the domain of life-sciences (HPC clusters vs. supercomputers).