WeNMR: an infrastructure for NMR and structural biology
Sara Coelho profiles the first research community to join EGI
WeNMR is a fast-growing e-Infrastructure community on the frontline of biomedical research, poised to become the first Virtual Research Community to sign a Memorandum of Understanding with the European Grid Infrastructure.
The structural biology community is already a rapidly expanding Virtual Organisation (VO) that accounts for about one in every six life science jobs submitted to the grid infrastructure. In 2010, this totalled around 670,000 jobs, corresponding to about four million CPU hours, on behalf of more than 250 users worldwide.
“Life sciences in general is too broad and the requirements very disparate,” says project coordinator Alexandre Bonvin, from the University of Utrecht. The agreement with EGI will allow WeNMR to represent the interests of its community better. “We also hope that recognition by EGI will facilitate the support to our VO from various NGIs [National Grid Infrastructures],” he adds. “This is important for us in terms of sustainability.”
The main goal of WeNMR is to offer a user-friendly e-Infrastructure for scientists and to establish a thriving virtual research community in the fields of nuclear magnetic resonance (NMR), small angle X-ray scattering (SAXS) and structural biology, building on the success of the previous eNMR project.
eNMR to WeNMR
While eNMR focused on taking the e-infrastructure off the ground, “now we are more into operating, consolidating and expanding this e-Infrastructure,” says Bonvin. The three-year WeNMR project will provide support to software developers and users and will promote the adoption of e-Infrastructures within the life sciences.
High-resolution NMR allows scientists to measure the distances between atoms of a molecule, which are then used to calculate the three-dimensional structure.
This is not just an academic exercise. Many high-impact diseases, such as Alzheimer’s or Parkinson’s, are caused by accumulation of protein-type molecules in the brain. Knowing their precise structures is the first step in the search for medical compounds able to defeat the disease. Calculating molecular structures out of NMR data requires a great deal of computing power and this is where a robust e-infrastructure can make a difference.
“Many applications are CPU intensive and some are rather complicated to install and maintain,” says Bonvin. The CS-Rosetta application, for instance, generates 3D models of proteins but requires 5,000 to 10,000 CPU hours to run. A local cluster of 100 cores would take five days to produce results. “On the grid this can be reduced to less than one day,” he adds.
The WeNMR e-Infrastructure uses EGI’s grid for the CPU intensive calculations. This will allow small research labs to benefit from state-of-the art methods even if they do not have adequate local infrastructure resources or expertise.
Bonvin believes that his community, which entered the grid business only rather recently, can bring a "fresh new look at things for the benefit of the wider EGI community.”
“We are following a strategy to provide easy and user-friendly access to the grid via web portals,” he says. “And we look forward to sharing our experiences and solutions with others in the cadre of the competence centre we are setting up within our VRC”.
