News April 07 Grid Project Ends

On Friday 27 April the Screensaver Project finally came to a close. The project, developed with the National Foundation for Cancer Research has run for six years and has at various times been funded by Intel, Microsoft and by IBM, but was chiefly a collaboration with United Devices Inc of Austin Texas.

It has been an enormous success, involving over 3.5 million personal computers in more than 200 countries. Only the SETI [Search for Extraterrestrial Intelligence] project has had more participants, but none has involved as much data transmission as this research.

The project built a database of billions of small drug-like molecules with known routes to synthesis. These compounds have been screened virtually to see if they might make potent inhibitors of proteins of known crystal structure and biological significance.

The chief targets have been proteins whose blockade might provide a mechanism to counteract cancer, but there were side projects aimed at drug leads against anthrax and against smallpox.

What has been achieved? Firstly on the methodological side, we have shown the viability of the virtual screening approach using a grid of personal computers. This has been adopted by a number of commercial companies.

In terms of actual predicted 'hit' molecules, those we have had synthesized and screened have shown biological activity: up to 10% of the predicted compounds being experimentally active in one specific case, which is high by industry standards. We still have many results to take forward for synthesis and testing and this is being done in collaboration with the National Foundation for Cancer Research. This aspect of the research will continue even though we are not going to be able to run new targets at present.

We would like to thank all the millions of volunteers who have given us their computer time: over 450,000 year's worth of it. Without them the project would have been impossible. With their help we have found novel leads which we hope will be taken forward for development.

Our heartfelt thanks to everyone.

News Jan 07

NFCR collaboration enters new phase

InhibOx and the NFCR are pleased to announce that the cancer screensaver project has entered a new phase.

The cancer screensaver project was a public interest endeavour that was started by the NFCR in 2001. The project uses the idle time of over 3.5 million computers to computationally screen a large database of molecular structures against 12 cancer targets. InhibOx has donated all of its right to the results of the project back to the NFCR.

As the priority for further processing, NFCR has provided additional research funding to support collaborative projects between Oxford University and InhibOx to analyze available data related to two of the twelve original cancer targets - CDK2 and FGF. The objective of these collaborations is to identify, with the technical assistance of InhibOx, a subset of several hundred molecules for further experimental testing. It is hoped that this work can be completed by the end of the first quarter of this year, enabling NFCR to start chemical synthesis and biological tests on these selected molecules for developing anticancer drug candidates. This is a significant milestone for the screensaver-lifesaver project.

News Dec 2005

Pancreatic Cancer project results presented at AACR-NCI-EORTC Molecular Targets Meeting Nov 2005

Our collaborators on the pancreatic cancer project: Vijay Gokhale, Haiyong Han, Daniel D Von Hoff and Laurence H Hurley working at the Arizona Cancer Center, Tucson, AZ and the Translational Genomics Research Institute, Phoenix AZ presented the results of the recent project.

"quote - Using Screensaver Lifesaver / LigandFit technology, we have identified three promising leads Mol597, Mol238 and Mol628 for urokinase inhibition. Based on the docking studies these molecules show high potential as uPA inhibitors. These molecules can be used as lead molecules for the design of better uPA inhibitors as potential anti-cancer therapeutic agents. We are synthesizing these molecules and testing them for biological activity. "

View the poster from the meeting. (requires Adobe Acrobat)

News 19 April 2005 Pancreatic Cancer Latest Target in Peer to Peer Computer Campaign

Cancer researchers running the world’s second largest virtual supercomputer announced today that their vast Peer-to-Peer Network has turned its attention to fighting pancreatic cancer.  In a joint statement, with the National Foundation for Cancer Research (NFCR), Dr. Graham Richards, Chairman of the Oxford University Chemistry Department and Director of the NFCR Centre for Computational Drug Discovery and Dr. Daniel Von Hoff, Director of the NFCR Center for Targeted Cancer Therapies at the University of Arizona in Tucson and the Translational Genomics Research Institute (TGen) in Phoenix, announced the new collaborative research project that will target developing cancer drugs to fight one of the world’s most deadly cancer types, pancreatic cancer.  Find out more

News 22 April 2004 Purely Proteins and Inhibox Launch Combined Drug Screening Programs

Purely Proteins Ltd (Cambridge, UK) and Inhibox Ltd (Oxford, UK) today announced the launch of commercial drug discovery services based on their combined proprietary computational and laboratory-based screening technologies.

This platform is commencing with the immediate launch of discovery-ready drug screening programs for families of human phosphatases, kinases and proteases. Key members of these families, such as PTP-1B, c-abl and beta-secretase, are proven therapeutic targets in life-threatening diseases such as diabetes, cancer and Alzheimer’s disease. Other therapeutic targets of immediate interest to clients can also be accessed on a fee-for-service basis.

The screening process involves the selection of highly specific virtual drugs against defined therapeutic targets from billions of putative starting compounds, using Inhibox’s proprietary computational processes. The resulting highly selected compounds are synthesized and tested in Purely Proteins’ comprehensive drug screening platforms. Experimental datasets are then made available for licensing either as new modules within Purely Proteins’ existing relational database product TargetBASE“, or as patented, early-stage licensing candidates.

TargetBASE already houses detailed structural, biochemical and pharmacological information about therapeutically relevant protein families. This new, experimentally exemplified chemical information will expand its content into the new field of chemical proteomics.

Commenting on the program, Purely Protein’s Chief Executive Officer, Dr David Bailey, said, “We are delighted to be working with Inhibox, a company with extensive experience in computer-based drug screening and computational chemistry. The combination of Purely Proteins’ expertise in purifying and assaying families of human proteins with the extreme power of Inhibox’s computational screening will add significant value to both companies’ products very rapidly.”

Dr Paul Finn, Chief Executive Officer of Inhibox, said, “Purely Proteins is an ideal partner for Inhibox. Our technologies have been developed to exploit large-scale virtual screening opportunities such as those presented by the human proteome, which we can reduce to practice through Purely Proteins’ extensive portfolio of discovery screens. Together, we can rapidly convert in silico designs into validated drug opportunities for new pharma and biotech clients. The collaboration with Purely Proteins is also an attractive mechanism for further progression of the results of the Internet-based Cancer Screensaver Initiative.”

Purely Proteins

Purely Proteins (PP) has been established to exploit the increasing need for human proteins in pharmaceutical research. The company purifies suites of human proteins, as both drug targets and bio-therapeutics, using proprietary processes that ensure both high purity and biological activity. PP also offers drug-screening services, backed by state-of-the-art informatics, to examine therapeutic targeting to these proteins. More details about PP, its products, and its screening services can be found at www.purelyproteins.com.

Inhibox

Inhibox, a recent spinout company from the Oxford University Chemistry
Department, specializes in automated computational approaches for virtual screening of compounds, using its molecular database comprising 35 million commercial and combinatorial library compounds and high-throughput grid computing based hardware platforms. In addition Inhibox is developing novel docking algorithms and offers broader computational chemistry services. Inhibox is a partner in the Internet-based Cancer Screensaver Initiative. Further information about Inhibox can be found at www.inhibox.com.

Contacts

News 12th December 2003 We are now nearing completion of Phase II of our in silico screening for anti-cancer drugs. Phase I, where we used the very efficient THINK software and screened a database of 3.5 billion possible drug molecules, produced a very large number of hits: far more than could be synthesized and tested. All of these targets have now been taken through Phase II, with the exception of superoxide dismutase, which proved to be unsuitable for further work.

Our brief excursions into seeking candidate lead compounds to provide protection against anthrax and later smallpox, taught us just how important it is to reduce the list to manageable proportions. In the anthrax case we provided too many unfocussed hits. In the case of smallpox, using the LIGANDFIT software, we produced some 900 hits, with a subset of about 50 of these representing really promising leads. We did this by evaluating several different methods for “ranking” the best hits, and prioritising those compounds that were identified as high-scoring by several methods. These lists were presented to representatives from the United States Department of Defense on 30th September 2003.

Developing methods for selection of the best set of hits for follow-up from the hit list has been the basis of Phase II of the cancer project and is drawing to a conclusion. We are close to having a reasonable number of novel suggestions for synthesis and testing. We envisage that the experimental results will help us to refine and improve these methods. Once this has been achieved we will need to engage industrial partners to perform the necessary and expensive development work, which will follow the discovery of genuine and novel drug candidates.

We have also added two new protein targets to keep active the wonderfully powerful system which the generosity of participants provides. This very week the number of computers that have joined the project has passed 2.5 million: an enormous contribution for which we are eternally grateful.

News 19th May 2003 In recent weeks UD.com have returned results back to Oxford from the members efforts. We have results from the first 6 cancer targets and the first binding site target in the smallpox project. The results data are currently being processed by Inhibox.

News 5th February 2003 Smallpox Project launched.

The Smallpox Project is running from the web site here:

http://www.chem.ox.ac.uk/smallpox

The threat posed by smallpox as a weapon of bioterror is widely perceived. Vaccination of huge numbers of people has many drawbacks such as a delay in efficacy and a percentage of damaging side effects. A drug to combat the effects of exposure to smallpox has many attractions.
The universally acclaimed screen saver project involving the collaboration between United Devices Inc and Oxford University's Chemistry department has demonstrated how massively distributed grid computing using the idle time on PCs around the world can be harnessed to screen many millions of molecules as potential drugs. The on-going project to discover anti-cancer drugs and the effort in the spring of 2002 to find drugs to block anthrax have demonstrated how grid computing can produce drug candidates in a very short time when utilizing the 2 million PCs which now make up the user community.

Now this combination with the generous support of IBM and new partners is launching a similar project to seek drug candidates for smallpox protection.

Screen saver time will be used to find small molecules which inhibit a key enzyme used by the virus to unpack its DNA. The target enzyme [topoisomerase] has a form present in all human cells where it is a target for anti-cancer drugs and the project will investigate both the virus and the normal targets.

New partners are Evotech OAI, the European biotechnology company who have prepared the input information on the protein target and Oxford University is being assisted by colleagues at Essex University who are part of Oxford's virtual research group, the NFCR Center for Computational Drug Design.The software used to estimate just how well a given molecule binds to the target, and hence just how good a drug it might prove has been provided by Accelrys Inc, its Ligandfit program, while smallpox expertise has been furnished by Drs Grant McFadden and Stuart Schuman.

Scientific coordinator Graham Richards of Oxford says
"Nothing could be more appropriate in seeking protection against a universal threat than to engage ordinary people across the world in a coordinated effort to discover a drug which would render smallpox impotent"

News 22nd January 2003 SCREEN SAVER 2 MILLION MILESTONE

Today the screen saver project run from the Department of Chemistry at the University of Oxford and United Devices of Austin Texas passed an impressive milestone. There are now over TWO MILLION personal computers involved in the project with almost one million individual people in 226 countries.

Launched in April 2001 the project, devised and lead from the NFCR Center for Computational Drug Discovery in the Chemistry Department headed by Prof Graham Richards, has had its main goal the discovery of molecules which could be candidates for anti-cancer drugs. Such has been the response from the general public that billions of molecules have been screened against protein targets which play a role in the development of cancer. In addition a year ago potential blockers of anthrax were sought, with those results given to the US and UK governments.

The hits from the ongoing cancer project are now being refined and it is hoped to start synthesis and testing very shortly.

The project has been a collaboration between Oxford's Chemistry Department and United Devices Inc who provide the grid computing technology. At various times the work has had sponsorship from Intel Corporation and Microsoft Inc.

Graham Richards said “This is one of the biggest computer projects ever undertaken and not only demonstrates the power of grid computing, but involves the general public in serious and important scientific research”

News 6th November 2002 A new research article has been published, the article was written half-way through the first phase stage of the project.

“Pattern recognition and massively distributed computing” E. Keith Davies, Meir Glick, Karl N. Harrison, W. Graham Richards Journal of Computational Chemistry Volume 23, Issue 16, 2002. Pages: 1544-1550

The article abstract is here:

A feature of Peter Kollman's research was his exploitation of the latest computational techniques to devise novel applications of the free energy perturbation method. He would certainly have seized upon the opportunities offered by massively distributed computing. Here we describe the use of over a million personal computers to perform virtual screening of 3.5 billion drug like molecules against protein targets by pharmacophore pattern matching, together with other applications of pattern recognition such as docking ligands without any a priori knowledge about the binding site location.

News 30 September 2002 All the data from UD.com from the phase 1 results have been collected and reported.

News 20 July 2002 First phase 1 results have been collected and reported.

Also Professor Richards has made two statements 1) about the Phase 2 of the project and 2) the exploitation of the results.

News 26 June 2002

An Update on the Cancer Research Project from United Devices

Thanks to the continued dedication of our project Members, the Intel-United Devices-NFCR Cancer Research project proved more successful then we could ever have anticipated. It was so successful that we will begin a second phase of the project shortly to continue this groundbreaking research. The first phase, using the THINK application to identify key "hits", is complete. The second phase will further refine these initial hits by running them through a different application called LigandFit.

United Devices is still working in partnership with the National Foundation for Cancer Research, Oxford University and Dr. Graham Richards. His team will continue to receive the results and work diligently to advance cancer research. During this transition time, Members will notice the new application running, evidenced by a new name and graphics. Otherwise, we are simply continuing the same research with the ultimate goal of finding more effective drugs with few side effects to fight cancer.

We would like to thank Keith Davies, author of THINK, for all his contributions to this project.

A word from Dr. Graham Richards on the science behind the project:

"The screen saver project has been an enormous success. It has involved well over a million and a half personal computers in more than two hundred countries, yielding in excess of 200,000 years of CPU time. With such power it has proved possible to screen three and a half billion molecules against a range of cancer targets and, in a three week burst, against the anthrax toxin.

Now we can look to the future and use this incredible power to extend the project using computationally more demanding but more reliable methods. The THINK software has the virtue of incredible speed with a very light demand on data transfer. We now intend to use the more sophisticated LigandFit software, an improvement made possible by virtue of the power that the user community has made available. We believe that the scoring function, which indicates just how good hits are likely to be, is more accurate than the crude but fast methods used hitherto. Our obvious next step is to take the lists of hits generated so far and to run them by the newer software so as to generate a smaller list of more accurately generated hits, thus leading to a more tractable set of potential drugs to synthesize and test."

News 28 May 2002

The members hit results for the protein targets 7-10 have been updated. The new results contain information from the E8 library of molecules and therefore adds some new molecules with good binding energies.

News 2 May 2002

The members hit results for the protein targets 7-10 are online. The molecules which scored lower than 1000 kJ/mol were returned as hits and the histograms shows the distribution of these crude hits as a function of the score. In a post-processing step, refinement of the position and conformation of the molecule in the receptor site is expected to decrease the scores of many of the molecules to below zero. Scores or free energy binding predictions below zero correspond to molecules which might be bind to and inhibit the protein receptor.

News. 19th Feb. 2002

The University of Oxford & United Devices is pleased to announce that as of 19th February 2002, the screening phase of the Anthrax Research Project has been completed, four weeks after the project was started.

News 1st Feb. 2002

Several visitors to the web site have asked how you can just run the cancer the anthrax project. This is simply done by opening the United Devices Agent software, clicking on the link which says "View your device list", this allows you to log-on to the Ud.com members web site. Then open "My Device Manager" and select "Profiles". Click on the profile you wish to edit and when it opens select either the "Intel-United Devices Cancer Research Project" or "United Devices Anthrax Research Project", then save your changes.

News 22 January 2002

Intel, Microsoft, United Devices, the Department of Chemistry at the University of Oxford and the National Foundation for Cancer Research (NFCR) ask computer users around the world to join with them in the Anthrax Research Project, an international effort designed to help scientists develop a treatment for the Anthrax toxin. More info here.......................

News 20 January 2002

The members hit results for the first 6 protein targets are online. The molecules which scored lower than 1000 kJ/mol were returned as hits and the histograms shows the distribution of these crude hits as a function of the score. In a post-processing step, refinement of the position and conformation of the molecule in the receptor site is expected to decrease the scores of many of the molecules to below zero. Scores or free energy binding predictions below zero correspond to molecules which might be bind to and inhibit the protein receptor.

News 19 January 2002

A new software package has been add to the web site to allow visitors to sketch their own molecules and check whether this molecule has good binding properties to the target proteins - and therefore you could find good anticancer candidates - if you do get a hit let us know!

News 23 December 2001

Professor Richards has donated a share stake in the newly formed company Inhibox.

Professor Richards, who lost his first wife to the disease and whose second wife has recently undergone surgery for breast cancer, has donated a 25 per cent stake in Inhibox, a new drug research business, to the National Foundation for Cancer Research.

Inhibox, which was made into a company on Friday, is a cancer drug discovery company that uses millions of individuals' home computers to match a huge library of molecules against targets on proteins associated with the disease. Find out more news on this story here from the online Telegraph, here from the online Guardian, or here from the Oxford Mail.

News 15 October 2001

Lots of good news.

Over one million computers are running the project and worked together for >55, 285 years of CPU Time.

Two new proteins will be the target from this week, they are Cyclo-oxygenase II (6COX) and BCR-ABL a tyrosine kinase (1IEP) which is believed to be causally involved with chronic myeloid leukaemia - see Science (2001) 292 p 399-400. See the protein pages for more info.

First results are in from UD.com.

The results pages show the resulted hits for the first two proteins. We will be starting here at Oxford to do the second phase screening (post-processing step) of these hits to get better targets for possible anticancer molecules.

News 2 August 2001

News of Cancer Project Stats: 834,273 computers working on the project, >33,000 (years) CPU Time. Also United Devices have created a new statistics reporting system for the Global MetaProcessor™ - which gives live data of the project and greatly improves the feedback to the members of the project.

News 16 July 2001

News of Cancer Project Stats: 788,273 computers working on the project, >28,317 (years) CPU Time

News 20 June 2001

News of Cancer Project Stats: 714,372 computers working on the project, 172,392,482 (hours) CPU Time

News 30 May 2001

News of Cancer Project Stats: 630,553 computers working on the project, 116,653,867 (hours) CPU Time

News 8 May 2001

The intellectual property statement for cancer project has been added to the web site.

News 2 May 2001

The University of Oxford project team have been very pleased with the response we are having to the project. It has been overwhelming and we have been looking at increasing the numbers of molecules, the precision of the conformational generation (i.e. the torsion angles considered), the numbers of targets and accelerating the timescales for their availability.

It was also apparent that the production version of THINK was running significantly faster than originally estimated. In addition, we have a new version (v1.11) in development which we hope to release this summer which runs much faster again. Part of the credit for these improvements should be attributed to INTEL because they identified some code bottlenecks which has allowed two important algorithms to be rewritten (The new algorithms use slightly more memory). Obviously, this new version needs to extensively validated prior to release.

There are a number of configuration options which affect how long each work unit takes including the number of molecules per work unit, the number of derivatives generated for each molecule and the precision of the conformational generation. At the time of the launch it was decided to gather some initial data which would allow us to improve the configuration of future work units. In addition, some statistics on the number of participants who don't return results, should address the question of whether sending out the same work unit repeatedly is necessary and whether this is especially true for some work units. I should also add that the preliminary results look very interesting from a scientific perspective with some molecules being predicted to bind quite well.

We hope to start using revised configurations this month with approximately 35 million drug-like molecules being downloaded plus 100 derivatives of each (i.e. 3.5 billion molecules in total). This will use a revised version of the software which will address a number of the concerns raised in feedback to date and analysing the results. These will mostly be at the minimum conformer precision, 200 molecules per work unit and 100 derivatives per work unit. After reviewing any work units which take abnormally long, we plan to repeat these at a higher conformer precision (probably with the new faster THINK v1.11 or v1.12). By doing so, we hope to learn a lot about improving the methodology as well as finding additional interesting molecules.

The initial 4 protein targets have been chosen and the current dates for the next 6 is mid-June and the final 6 by the end of Q3. The response to the project allows us to consider variants of some targets e.g. a different site on the protein or different conformation of the receptor site. We would prefer to focus on targets which have a good therapeutic potential, so no-one should be surprised if you see the same target come around again as a variant. In addition, with the increase in the numbers of molecules being processed we are considering distributing some of the follow-up calculations e.g. refining the geometry of a hit molecule in the receptor site.

In the future, we will report statistically summaries of the progress in scientific terms. This will include the total number of hits and numbers of molecules processed. If possible we will expand on this (and are open to suggestions what additional information to provide).

Figure. A view of the Linux cluster in Professor Richards' research group. The Linux cluster was in part funded by the NFCR and shall be used to look at the results from the cancer screening project. There are currently 45 nodes in the cluster running PBS queuing software to control the computational calculations.