WISDOM: A Grid-Enabled Drug Discovery Initiative against Malaria - Grid Computing: Infrastructure, Service, and Applications - page 370

1% 2%

2%

3%

3%

3%

EGEE Germany Switzerland
EGEE Asia Pacific
EGEE Russia
AuverGrid
EuChinaGRID
EELA
EGEE South Western Europe
EGEE Central Europe
EGEE Northern Europe
EGEE Italy
EGEE South Eastern Europe
EGEE France
EGEE UKI

3%

5%

38%

6%

7%

12%

15%

FIGURE 14.6 Distribution of jobs on the different grid federations.

available at this time. For WISDOM-II, we started the deployment with
only one server as well. Very soon, up to three servers were made available
at the SCAI Fraunhofer institute (http://www.scai.fraunhofer.de), with
3000 licenses available on each server. The FlexX software binaries were
stored like all the inputs on the grid storage elements and were installed on
the l y on each worker node at the beginning of the job.
As the average duration of a job was around 30 hours, we submitted one
instance per day, with a delay of 30 seconds between each submission. As
one instance was submitted in about 20 hours, the submission process
was quite continuous during the i rst month of deployment. The jobs were
submitted to 15 resource brokers in a round-robin order. At the end of a
job, the results were stored on the grid storage elements, and directly into
a relational database.
Figure 14.6 shows the repartition of jobs on the different grid federa-
tions. It also shows the AuverGrid, EuChinaGRID, and EELA infrastruc-
tures contribution. Each of these three infrastructures ran one single
instance that corresponds to 3% of the total 32 instances.

14.4.4 Postdocking Analysis and MD Refinement

Analysis of the docking results was carried out on the four targets docked.
The University of Modena took charge and analyzed the results obtained
on DHFR targets from P. vivax and P. falciparum . CSIR, in collaboration
with LPC Clermont-Ferrand, analyzed results on GST. Two criteria were
employed for result analysis: statistical analysis based on scoring values
and analysis of compounds making interactions to key residues of the
receptor. Based on these two criteria, the best compounds coming out of
the docking step were selected and rei ned using the MD procedure
described previously [15].