OpenMolGRID

OpenMolGRID

http://www.openmolgrid.org/

Molecular modelling, molecular engineering and drug discovery, as closely related fields, provided the set of real life applications central to the OpenMolGRID project. OpenMolGRID was conceived to exploit the power of Grid Computing to shorten the time to solution for drug discovery, specifically the identification of promising new compounds as potential drug candidates. The project demonstrated that the automation of the molecular design pipeline has promising advantages with respect to the manual, stepwise approach. It is much faster and more reliable, offering new tools. This will broaden horizons and open up new challenges related to the larger sets of potential promising compounds.

Challenges

The main workflow-related challenges in OpenMolGRID were to:

• simplify and automate the mapping of complex scientific workflows onto a set of Grid sites
• automate the resource selection process
• make better use of the available compute capacity in the Grid by splitting data-parallel tasks
• assist the user by automating low-level tasks such as inserting data transfers, choosing names for intermediate files, etc.

The workflows in question consisted of a sequence of database operations, data conversions and computational steps. The targeted end users were chemists, with a wide variety of IT experiences, from the "point-and-click" type to experienced developers.