Dears, we are happy to announce that the results of various GPUGRID-enabled studies were recently published in high-impact journals.
The first one I'd like to mention is "A high-throughput steered molecular dynamics study on the free energy profile of ion permeation through gramicidin A". The paper is a numerical investigation on the convergence of a molecular dynamics (MD) technique called "steered MD" (SMD), which is used to obtain free energy profiles (i.e. probabilities of occurrence of certain biological events). The SMD technique is essentially based on statistics: one has to repeat the same experiment (computational or otherwise) over and over, many times, until a reliable result is obtained. How many times, and how reliable? Up to now, studies performed SMD on biologically-realistic systems were limited to just a few (~10) trajectories, and the corresponding error was therefore hard to assess. We used GPUGRID to perform increasing number of pulls, up to two thousands, to verify how much the precision of a free energy profile was gradually improved as a consequence. (The answer, if you were wondering, is that one needs quite a lot of simulations).
T. Giorgino and G. De Fabritiis, A high-throughput steered molecular dynamics study on the free energy profile of ion permeation through gramicidin A, in press J. Chem. Theory Comput. (2011).
Thanks to all of the crunchers of the GA* series of workunits!