Tapping World Community Grid, the Help Conquer Cancer Project created a system that accurately recognizes when protein samples undergo a solidifying process called crystallization, which makes the proteins ready for further examination by special x-ray. The process is necessary for identifying, and eventually exploring, how the structure, shape and interaction of some proteins may have a role in causing cancer.
Using the Grid, scientists trained the system to successfully recognize 80 percent of crystal-bearing images and 98 percent of the clear drops of protein solution that existed prior to crystallization. This enables six times as many images per protein to be examined compared to human review, and in dramatically less time.
Automating the identification of crystals could speed research in numerous biological science and genetic research projects, as crystallization holds the key for investigating a variety of biological processes. It could also validate the efforts of other projects that seek to obtain protein structures, such as the Nutritious Rice for the World effort, which is also using the Grid to explore ways to create hardier, healthier strains of rice.
An article about the breakthrough, authored by the lead researchers of the Help Conquer Cancer Project, was recently published in the Journal of Structural and Functional Genomics (http://www.springerlink.com/content/750430466h336142/fulltext.pdf).
Crystallization allows beams of x-rays to bombard the sample and diffract light into many directions, producing a three-dimensional profile, making samples easier to study. However, the crystallization process is tedious: A single protein sample may require thousands of attempts to spur crystallization by introducing chemical compounds via robot. And, until now, painstaking human observation was then still needed to verify that crystals actually formed.
X-ray crystallography has played a key role in many scientific endeavours. For example, it was used in the mid-twentieth century to help determine the shape of DNA. Proteins aren't the only substances that can form crystals for easier study and development; scientists use x-ray crystallography on metals, minerals and semiconductors as well.
"This advance illustrates once again the enormous value that World Community Grid brings to the scientific arena", stated Dr. Igor Jurisica, senior scientist at the Ontario Cancer Institute, associate professor in the departments of computer science and medical biophysics at University of Toronto, visiting scientist at IBM's Centre for Advanced Studies, and Canada Research Chair in integrative computational biology. "The people who volunteered the spare computing cycles of their PC processors ought to take great pride that they made this development possible."
"Advances such as this make us at IBM feel privileged to be in a position to support World Community Grid and the work of the many scientist who use the system", stated Dr. Joseph Jasinski, Distinguished Engineer and Programme Director of IBM's Healthcare and Life Sciences Institute. "We're confident that the advance announced today will prove to be a valuable contribution to the ways in we better understand disease."
Cancer caused about 13 percent of all human deaths in 2007, according to the World Health Organization. The group says that there were 12,7 new cancer cases and 7,6 million cancer deaths in 2008. In 20 years, the number of new cases diagnosed annually will soar to 21 million, with 13 million cancer deaths.
The Help Conquer Cancer project was launched on the Grid on November 1, 2007, under the auspices of the Ontario Cancer Institute at Princess Margaret Hospital at University Health Network in Toronto, Canada, and Buffalo, New York's Hauptman-Woodward Medical Research Institute.
Since then, Grid volunteers have contributed 50.981 CPU-years to the cancer project to date, an average of 54 years of computing per day. It is helping to identify and map Hauptman-Woodward's archive of 100-million images of 12.500 unique proteins that could be linked to cancer, captured in the course of more than 19,2 million experiments there.
This comprises the most comprehensive database on the chemistry of a large number of proteins, a resource that will help researchers around the world unlock the mystery of how many cancers grow, such as breast, prostate or childhood leukemia.
Computer examination provides both qualitative and quantitative advantages. For instance, it is impractical for humans to review all 9216 images for a given protein. And even if a person tried assessing one image per second, it would take 1333 days to examine all 12.500 proteins under study. Human evaluations may also vary widely and reveal inconsistencies, even from the same person.
In earlier stages of automating the process of verifying crystallization, computers were accurate only about 70 percent of the time, and could examine only about 850 features, compared to about 15.000 now capable of being analyzed.
Help Conquer Cancer runs on World Community Grid, sponsored by IBM. By aggregating the unused cycle time of 1,5 million personal computers donated by hundreds of thousands of volunteers in more than 80 countries, World Community Grid is the world's largest public humanitarian Grid, equivalent in strength to one of the world's most powerful supercomputers. More than 22.000 new devices were added to the Grid in April 2010 alone.
World Community Grid provides the equivalent of about 400 teraFLOPS of speed, or 400 trillion floating-point operations per second. The Help Conquer Cancer application performs about 80 trillion floating-point operations per second.
Other projects running on World Community Grid have produced results including:
- Discovering Dengue Drugs Together project has identified potential antiviral compounds to fight Dengue Fever and related illnesses.
- Help Fight Childhood Cancer finds drugs that can disable three proteins associated with neuroblastoma, to make one of the most frequently occurring solid tumours in children more curable.
- Researchers at Scripps Research Institute working on the FightAIDS@Home project recently found two compounds that will lead to new treatments for drug-resistant HIV strains.
- Nutritious Rice for the World project has completed 12 million computing transactions over 11.000 years of compute time in an effort designed to yield hardier, healthier strains of rice.
- An effort to help cure muscular dystrophy and other neuromuscular diseases.