Find-a-Drug   Issue 2 13 September 2002

Ready, Steady, Go ...

Following successful alpha and beta test phases, Find-a-Drug began running the first gold, or production, version of the software towards the end of July. Calculations also started on a number of new protein targets, and the number of molecules being processed continues to be increased.

The Cancer project is a continuation of the work started with United Devices. The initial focus is on protein targets which could not be fully evaluated with the original version of THINK. This includes VEGF (Vascular Endothelial Growth Factor) receptor which controls the growth of blood vessels. Another interesting therapeutic possibility is to find an inhibitor to HDM2 (Human Double Minute 2) protein - this is a protein which prevents the body's natural defences killing cancer cells.

The Methodology project focuses on protein targets which can help to improve the software. All members participate in this project as well as one or more of the other projects that they select in the Find-a-Drug Control Panel. The Methodology project is also used to test the suitability of the software for a new project by running a set of molecules on a protein target.

It is hoped to start the Bioterrorism Antidotes and HIV projects later this year. Find-a-Drug has also received some interesting collaboration proposals and it is hoped that future projects might address diseases such as Multiple Sclerosis.


What are project statistics?

The project statistics pages are intended to provide detailed information on the progress of each protein. An analysis of the number of hits, their scores, the time taken per job and the level of repetition is also included. We publish this information as part of our open policy as it can help members to understand their contributions. It is updated every 6 hours - the same frequency as the member, country and team statistics are updated.

Job repetition
Results per job  Percentage of jobs
5+   21%
4    12%
3    15%
2    22%
1    30%
0    0%
Hit scores
Score  Number of hits
> -20     551
> -40     563,354
> -60     4,951
> -80     9
> -100    0
<=-100   0
Job hits
Number of hits  Number of jobs
>10   7,053
9-10   402
7-8    406
5-6    316
3-4    260
1-2    606
0      18,188
Job times
Time (hours)  Number of jobs
>24   2,241
<24   1,604
<16   3,614
<8    4,414
<4    6,145
<2    5,935
<1    3,272

The term redundancy is sometimes used in place of repetition and often needs explanation. Given the nature of internet-based computing, there is scope for different results for the same job. Small discrepancies do arise from rounding behaviour with different processor types but there have also been examples with other projects when results have been fabricated. It is also inevitable that some jobs are abandoned with the consequence that not all dispatched jobs return results. At Find-a-Drug we endeavour to run at a minimum level of redundancy which is substantially lower than that used by similar projects (including the first distributed computing project using THINK run by United Devices). By using a low level of redundancy we are able to make better use of members' PCs and process more molecules for more protein targets.

Some feedback

"Nice and stable."
Nickname: The BFG

"I'm a FaD user since beginning and there has been major progress on this project. Really astonishing ..."
Nickname: Supp


How can you help?

Word of mouth is probably the most effectively way of promoting a distributed computing project. So if you are running Find-a-Drug tell others, if you are not then tell us why not! You can also just download the latest version and discover the improvements for yourself.

If you experience problems running Find-a-Drug contact us via e-mail, use our feedback form or post a query on our message board.

Find-a-Drug is a tradename of Treweren Consultants Ltd Copyright © Treweren Consultants Ltd 2002