TORONTO – There’s something to be said for case studies that tell you how to do something successfully.
But the value of failure is highly underrated, says Igor Jurisica, a scientist at the Ontario Cancer Institute in Princess Margaret Hospital’s division of cancer informatics.
speaking Monday at a briefing on how information technology is helping advance the University Health Network’s research on protein crystallization, points out that it’s the nature of publishing to feature only the success stories.
“”If a molecule has been crystallized, you have a short-hand recipe of how the molecule was crystallized,”” he says, comparing the process to cooking. “”You don’t have pages and pages saying, ‘We tried for the last five years and here are all the recipes that didn’t work,'”” he says. “”I would like to know what else failed before I repeat the failure myself.””
Jurisica explains that protein crystallization (or crystallography) is important because it allows researchers to determine the three-dimensional shape of a protein, which can then be used to determine how the protein interacts with other molecules and ultimately to develop drugs to fight diseases.
“”No single gene is responsible for most diseases,”” says Jurisica. “”It’s a combination of various genes and molecules interacting. We need sophisticated analysis to understand this network of interactions.””
Using IBM Corp.’s RS6000 scalable server architecture and DB2, the vendor’s Web-enabled relational database management system, Jurisica’s lab has developed a computerized method of subjecting proteins to 1,534 variable conditions. In the past, he says, crystallography was determined using only 90 variables and the work was done manually, using human – read subjective – researchers.
Now, using powerful imaging technology and robotics, researchers can see a greater number of and more precise results.
DB2, he says, also answers another challenge that biology research has traditionally faced.
Until recently, biologists didn’t use computers in their research. Instead, many experiments were recorded in shelves of books lining labs everywhere, making it impossible for other researchers to access that knowledge.
Biologists, however, are starting to realize the potential IT offers to speed disease research and drug development, says Jurisica.
“”There are so many different pieces of the puzzle around the world in different formats it would be foolish not to use them – but at the same time it’s very difficult to use those pieces.””
The UHN, adds Jurisica, is in the process of moving to IBM’s DiscoveryLink system. This comprises DB2 software, DB2 Relational Connect middleware tailored to life sciences research, and DB2 Life Sciences Connect, which IBM says enables a DB2 federated system to integrate a wide variety of research data from heterogeneous sources.
While the DiscoveryLink system won’t necessarily mean Jurisica will be able to do things he’s not doing today, “”to some degree it will enable us to do things more smoothly,”” he says.
According to Sharon Nunes, director of solution development for Life Sciences at IBM, the merging of information technology and biology research is driving an enormous explosion of data.
Five or six years ago, she says, drug companies were investigating about 500 targets to develop drugs against.
With the recent publishing of the human genome, that number has swelled to 5,000 to 10,000.
A three-year-old genomics company has more than 250 TB of data stored – and that’s not counting the data that’s regularly discarded, Nunes notes.
And many public databases are doubling the amount of data they hold every six to eight months, she adds.
“”This means we need the systems and software to manage that data.””
At the same time, the cost of developing drugs has increased by about five times what it cost some 25 years ago – to US$800 million over 12-15 years to develop a current drug on the market, according to one study – and the number of drugs that has come out of the research has remained the same, Nunes says.
“”Why aren’t we seeing the number of drugs coming out of the other end if there’s this improvement in input? If you look at the amount of data being generated, one of the problems is we’re creating bottlenecks.””
Nunes cites silos of information, multiple data types and the lack of standards as the main challenges facing the biotechnology and pharmaceutical research industries.
“”The succesful biotechnology companies are going to be the ones who figure out how to use IT to their advantage,”” she says.
UHN expects to complete its move to DiscoveryLink within a few weeks.