[BOGOTÀ] Colombian scientists are using a global network of personal and institutional computers to search for potential drugs against leishmaniasis, a disease that affects 12 million people worldwide.
They will harness the calculation potential of the almost two million computers that make up the World Community Grid, funded by the IBM Corporation.
The project will use an existing open-access database of 13 million drugs to find those that can destroy or inactivate some of the leishmaniasis parasite's 53 proteins. Scientists will concentrate on 600,000 candidates, which have been shortlisted.
Carlos Muskus, the project leader at the University of Antioquia, Colombia, told SciDev.Net that, thanks to the grid, 100 years-worth of work can now be carried out in less than three years.
Muskus said that promising drug candidates would undergo a series of tests, before clinical trials and eventual release as a treatment.
Mauricio Hernandez, a bioinformatics expert at the same university, said the work is pioneering in Latin America. He added: "With this methodology scientists in developing countries can obtain for free what would otherwise cost US$3–4 million — the price of a supercomputer with 2,000–3,000 processors".
The World Community Grid is searching for drugs for other diseases, including dengue and HIV/AIDS. Anyone can join, and members of the community can choose the project they want to support.
There are several similar initiatives around, such as Wide In Silico Docking On Malaria (WISDOM), which uses grid computing to analyze the proteins of the malaria parasite.
Mauricio Rodriguez, director of the National Biotechnology Program and coordinator of the Bioinformatics Center that is being created in Colombia, said that computational biology is enabling researchers to "set [their] foot on the accelerator of knowledge", especially in developing countries.
But Stanley Watowich, an associate professor at the University of Texas, United States, who was involved with the grid's dengue drug search, said that "the process does have its drawbacks".
Watowich said that only 2–5 per cent of the network's users are willing to let their computers conduct large calculations that can slow down their machines. It is mainly the involvement of clusters at universities and institutes that can take the workload.
"We need to work out how to divide these calculations into smaller pieces that will not slow individual computers down too much."
Watowich added that the current process focuses on looking at the interactions between two proteins. This can be processed by a computer in a matter of minutes, but he warned that it is an over-simplification of the interaction and can lead to the identification of many unsuitable drug candidates.
"The current process works well but it also gives a lot of false positives," he said.
Additional reporting by Jan Piotrowski.