In 1985, the Search for Extraterrestrial Intelligence began scanning the night sky for radio signals from distant planets. This was only the first step in the search for intelligent life in the universe. Little did SETI know it then, but it would also be the impetus to discover a revolutionary new way to use computers to process data in the field of bioinformatics.
And bioinformatics, according to Boston University faculty and administrators, is one of the most exciting things happening on campus right now.
It all began when the endless streams of recorded data SETI collected in its search had to be analyzed somehow.
The most appropriate choice to crunch all that data: supercomputers. Supercomputers, though, are expensive, and SETI looked for more cost-effective alternatives.
What could be cheaper than the internet?
By harnessing the processing power of computers connected to the internet, members of SETI realized they could, in essence, create a supercomputer.
In 1996, project SERENDIP, the Search for Extraterrestrial Radio Emissions from Nearby Developed Intelligent Populations, began at the University of California at Berkley. Thousands of people volunteered their computers to aid in the search for alien life the first and most successful example of what is now called distributed computing.
In an era when every new computer model seems more complicated, distributed computing employs simplicity. Anyone with internet access can visit the web site of a distributed computing project like SETI. One can download a simple program that allows his or her computer, called a client, to connect with a host computer at the research facility where the project is managed.
Data, such as millions of radio emissions, is broken up into smaller parts, or work units, that can be more easily calculated. The host computer sends these client computers work units to process. After a period of time, the client contacts its host and sends it the completed work unit.
Richard P. Howell IV, a retired computer engineer living in Waltham, Massachusetts, participates in another distributed computing endeavor called [email protected].
‘I came across it by accident. I was browsing the web, checking my Intel stock and saw an ad for it,’ said Howell.
Howell and others donate their computers’ processing power to ‘helping us learn how living processes work,’ according to Dr. Vajda Pande, a postdoctoral student at Berkley.
Pande found inspiration in SETI’s project and started [email protected] two years ago at Stanford University to unravel the intricacies of protein folding.
Protein folding is ‘computational expensive,’ says Dr. Stefan Larson, lead researcher for [email protected]. ‘It takes a lot of computer time. You have to simulate molecules and their interactions at the atomic level over a long period of time.’
Understanding protein folding is necessary for drug discovery and disease progression. ‘There are proteins in the body folding and unfolding right now,’ said Chris Snow, a graduate student working on the project at Stanford. ‘It’s a fundamental part of biology that we need to know.’
Howell’s computer, and 90,000 others, donates unused computer cycles to [email protected]. A computer’s processing power is divided into cycles, with certain tasks requiring more cycles than others. Cycles go unused when a computer is sitting idly, or running a program that does not require much power. Unused cycles are usually lost, but distributed computer projects harness these unused cycles to complete their work units.
With [email protected]’s program, a window opens on the donated computer’s desktop, showing the molecule the computer is simulating. The program is relatively simple, rarely interfering with the computer’s daily uses.
‘The program’s goal is not to cause problems [with your computer],’ said Howell, adding, ‘you sort of never even notice that it’s there.’
Donated computer cycles have a dramatic impact on the project’s progress on understanding molecule reactions. ‘There are many protein malfunctions that have to do with incomplete or mis-folded proteins, such as Alzheimer’s disease,’ said Snow. ‘This project can help us understand how they go wrong.’
[email protected] is one example in the growing field of bioinformatics, the combination of biology and computer science. The Human Genome Project and others like it continue to generate masses of biological data.
‘The sophisticated techniques from computer science can be applied to analyze the huge amounts of data flowing from advances in genetics, genomics, and proteomics,’ said BU Provost Dr. Dennis Berkey.
Many believe this will have a significant impact in medicine and health care, including Dr. Charles DeLisi, director of BU’s bioinformatics program and a pioneer in the Human Genome Project.
DeLisi explained that the computer data analysis techniques are used to compare the gene expression in normal and cancerous cells to develop more appropriate drugs.
‘The inside of a cell looks like a circuit board that’s adaptable. We need to understand that connectivity better and once we do we’ll be able to reverse engineer [the cell and] have drug targets,’ said DeLisi.
Bioinformatics is still in its infancy. Just last year BU’s program, founded in 1999, boasted its first graduate. Currently, 90 masters and doctoral students are enrolled in the program.
‘The majority of students fall into two classes,’ said Dr. Simon Kasif, Boston University professor of computer science. ‘They just completed their BA or MA or they decided to spend a couple of years in industry and came back [to get a degree].’
‘Fortunately, BU has a wonderfully open academic environment, a spirit of collaboration among faculty from different departments and schools and outstanding scientific leaders,’ Berkey said.
A unique dual mentoring program reinforces this collaboration. Doctoral students have one computational and one wet lab advisor. ‘In many ways this goes against old academic nature. An advisor typically generates students in his own mold. Two advisors have different approaches. But in my lab, this is working very well,’ said Kasif.
‘BU recruits students from a variety of different fields and backgrounds. This makes the student population very diverse and students can help each other to fill in the gaps in each others’ backgrounds, which is crucial in such a multidisciplinary field,’ said Julian Mintseris, a postdoctoral student in the program.
BU’s bioinformatics program will continue to grow as the field does, according to Berkey, and it is only the beginning.
‘There is no question in my mind that this century is known as the century of biology,’ said Dr. Oscar Garcia, a bioinformatics expert at Wright State University in Dayton, Ohio. ‘It used to take forever to get a genome. Now we’re swimming in a sea of data … right now we’re breaking new ground.’