GRIMD: distributed computing for chemists and biologists

Motivation: Biologists and chemists are facing problems of high computational complexity that require the use of severalcomputers organized in clusters or in specialized grids. Examples of such problems can be found in molecular dynamics (MD), insilico screening, and genome analysis. Grid Computing and Cloud Computing are becoming prevalent mainly because of theircompetitive performance/cost ratio. Regrettably, the diffusion of Grid Computing is strongly limited because two main limitations:it is confined to scientists with strong Computer Science background and the analyses of the large amount of data produced can becumbersome it. We have developed a package named GRIMD to provide an easy and flexible implementation of distributedcomputing for the Bioinformatics community. GRIMD is very easy to install and maintain, and it does not require any specificComputer Science skill. Moreover, permits preliminary analysis on the distributed machines to reduce the amount of data totransfer. GRIMD is very flexible because it shields the typical computational biologist from the need to write specific code for taskssuch as molecular dynamics or docking calculations. Furthermore, it permits an efficient use of GPU cards whenever is possible.GRIMD calculations scale almost linearly and, therefore, permits to exploit efficiently each machine in the network. Here, weprovide few examples of grid computing in computational biology (MD and docking) and bioinformatics (proteome analysis).

Availability

GRIMD is available for free for noncommercial research at www.yadamp.unisa.it/grimd

Supplementary information

www.yadamp.unisa.it/grimd/howto.aspx