The following applications have been enabled on TCG@NUS:

Name	Description
AutoDock	Tool for automated docking of flexible Ligands to Macromolecules
BLAST	Tool for comparing nucleotide or protein sequences to databases & calculating the statistical significance of matches
HMMER	A sequence analysis tool using profile hidden Markov models
MODELLER	A program for comparative protein structure modeling by satisfaction of spatial restraints
Parametric	A computer program to run many times with different input parameters
PovRay	A high-quality tool for creating stunning 3D graphics

Applications Performance on TCG@NUS Grid

The Cycle Harvesting Grid is most suitable for running data parallel and parametric study type of applications. Application areas include the Life Science/Bioinformatics Research, Financial/Statistical Computation and Digital/Image Processing and other Computational Science and Engineering research.

1. Image Rendering/Processing
• Application software: POVRay
• Render a 1024x768 pixel image by Mark Slone
• Took one day and 20 hours to complete on a Dell 3GHz PC.
• Took about 100 mins to complete on 100 computers in TCG@NUS.
• Speedup: About 20 times faster.


2. Comparative Modelling of Protein Structures
• Application software: Modeller
• Run 9999 jobs consisting of protein sequences with about 200 amino acid residues
• Took 42 days to complete on a 600MHz PC.
• Took about 3 days to complete on 100 computers in TCG@NUS.
• Speedup: About 14 times faster.


3. Protein Level Sequence Matching
• Application software: tblast
• 22K sequences have to be searched against 3,461,799 sequences
• Estimated to take about a month to complete on a 16-CPU server.
• Took 20.5 hrs compute time on TCG@NUS and 13 hrs reassembly time on a standalone PC.
• Speedup: About 20 times (320 times with reference to a single CPU).


4. Large Scale High Accuracy Queue Simulation
• Application software: own code
• High accuracy numerical analysis for typical queues modeled as Markov chgins with infinite states.
• Estimated to take about 100 days to complete on a desktop computer.
• Took a few days compute time on TCG@NUS with about 500 computers connected actively.
• Speedup: About 20 times (312 times for some jobs).

Grid Enabled Application Software

Besides POVRay (http://www.povray.org/), Blast (http://blast.ncbi.nlm.nih.gov/Blast.cgi), Hmmer (http://hmmer.wustl.edu/) and AutoDock (http://www.scripps.edu/mb/olson/doc/autodock) other software packages that have been Grid-enabled on the Univa UD Grid MP Platform (but have not been installed on the TCG@NUS) include:

• Gold: Drug discovery - protein docking, to identify drug candidates.(http://www.ccdc.cam.ac.uk/products/life_sciences/gold/)
• Modeller: Bioinformatics - comparative protein structure modeling by satisfaction of spatial restraints, used for homology or comparative modeling of protein 3D structures. (http://salilab.org/modeller/)
• LigandFit: Drug discovery - protein docking. (http://www.grid.org/projects/cancer/ligandfit.htm)
• FlexX: Drug discovery - for predicting receptor-ligand interactions. (http://www.biosolveit.de/FlexX/)
• Archimedes: Simulation tool for healthcare - ?virtual patients?and drug effects. (http://www.kpcmi.org/media/archimedesqa.html)
• FRED: Drug discovery. (http://www.eyesopen.com/products/applications/fred.html)
• Glide: Drug discovery. (http://www.schrodinger.com/Products/glide.html)
• Amber: Drug discovery - molecular modeling and simulation. (http://amber.scripps.edu/)
• Rosetta: Bioinformatics - genomic sequencer. (http://www.rosettabio.com/)
• Winbugs: Bayesian inference on complex statistical problems. (http://www.mrc-bsu.cam.ac.uk/bugs/winbugs/contents.shtml)
• Excel: Perform analytics - Monte Carlo simulations.
  
  

NUS researchers are welcome to propose the application software they would like to run on the Cycle Harvesting Grid. We will provide the resources to help Grid-enable both the suitable open-source and in-house developed applications. Please feel free to contact us at .