Like every bubble, this one had to burst. Stock prices of many bioinformatics firms have fallen sharply in recent years. LION Biosciences of Germany went public at $40 a share and now trades at about $13. Iceland’s DeCode is worth a fourth of its former high. Even Celera, the U.S. firm that helped decode the human genome, is off its peak.
Falling stock prices are a symptom of a greater disappointment in bioinformatics. A few years ago the laborious and quirky process of drug discovery seemed on the verge of giving way to new streamlined, data-driven methods. Some firms organized the blizzard of genetic data into databases that researchers could mine with search engines from still other firms. Software companies built computer programs that modeled what goes on in human cells and even whole organs. Many investors came to believe that bioinformatics would open a new avenue to the discovery of drugs. But this avenue simply hasn’t materialized. Says biotech analyst Earling Refsum at Nomura Bank in London: “Bioinformatics has not helped Big Pharma get more drugs into the pipeline.”
Computers have succeeded in speeding up drug development, but this does nothing to address the far more vexing problem of finding drugs to treat new types of diseases. The obesity drug P57, currently under development, is a case in point. Researchers are now using molecular databases to screen new chemicals that might reproduce the appetite-suppressing effect of a natural plant. But the idea for the drug came not from computers but from tales of the San bushmen in South Africa, who for years have chewed a local plant to stave off hunger and thirst. Likewise, Swiss pharmaceutical firm Novartis used computer modeling to speed up the design of its leukemia drug Gleevec, introduced last year, but the original idea for the drug came from doctors who, back in the 1960s, noticed that all patients with the disease carried the same abnormality in their chromosomes.
Part of the problem may be that bioinformatics firms haven’t appreciated the idiosyncratic nature of drug discovery. Vertex had so much trouble with computer modeling software programs developed by outside firms that it decided to develop its own. Many bioinformatics firms are now trying to get closer to the drugmakers, either by developing in-house expertise in drug research or by partnering with Big Pharma. DeCode’s original business plan was to peddle its database of genetic information on the Icelandic population, but recently it began partnering with Roche in developing new drugs.
What is the endgame of all this partnering? Big Pharma may wind up absorbing the nascent bioinformatics industry into its own corporate research departments. Since drug development is data intensive, drug companies already have big information-technology departments. Existing companies may either try to turn themselves into drug developers, or become the domain of computer companies such as IBM, H-P, Compaq and Oracle, which are developing new computer technologies as fast as drug companies can buy them.
So far, the marriage of biology and computers has yielded “grid-computing,” a technology for putting a loose network of thousands of computers to work on a single big problem. Biotech firms have been pushing this method as an affordable alternative to buying bigger and bigger supercomputers. Eventually, computer “grids” may provide number-crunching power in the same way utility grids provide electrical power. “Biology is now the driver for high-performance supercomputing,” says Carol Kovac, director of IBM’s Life Science Division. Stephen McGarry, an analyst for Goldman Sachs, thinks there’s a good chance that the biotech and computer companies may eventually ally with one another, with computer giants taking a cut of drug royalties. If this comes to pass, perhaps bioinformatics experts by then will have succeeded at last in building computer systems that can discover new drugs. But right now, it’s not at all clear that they will.
title: “The Gene Bubble” ShowToc: true date: “2023-01-20” author: “Rosalie Andersen”
Like every bubble, this one had to burst. Share prices of many bioinformatics firms have fallen sharply. LION Biosciences of Germany went public at $40 a share and now trades at about $12. Iceland’s DeCode is worth a fifth of its former high. Even Celera, the U.S. firm that helped decode the human genome, is off its peak of $276, now trading at about $20. Many investors came to believe that bioinformatics would open a new avenue to the discovery of drugs. It hasn’t materialized. Says biotech analyst Earling Refsum at Nomura Bank in London: “Bioinformatics has not helped Big Pharma get more drugs into the pipeline.”
Computers have succeeded in speeding up drug development. But this does nothing to address the vexing problem of finding drugs to treat new types of diseases. Consider the obesity drug P57, currently under development. Researchers are using molecular databases to screen new chemicals that might reproduce the appetite-suppressing effect of a natural plant. But the idea for the drug came not from computers but from tales of the San bushmen in South Africa, who for years have chewed a local plant to stave off hunger and thirst.
Bioinformatics firms may not appreciate the idiosyncratic nature of drug discovery. Vertex had so much trouble with computer-modeling software programs developed by outside firms that it decided to develop its own. Many bioinformatics firms are now trying to get closer to the drugmakers, either by developing in-house expertise in drug research or by partnering with Big Pharma. DeCode’s original business plan was to peddle its database of genetic information on Icelanders, but recently it began partnering with Roche in developing new drugs.
What is the endgame of all this partnering? Big Pharma may wind up absorbing the nascent bioinformatics industry into its own corporate research departments. Since drug development is data intensive, drug companies already have big information-technology departments. Existing companies may either try to turn themselves into drug developers, or become the domain of computer companies such as IBM, H-P, Compaq and Oracle, which are developing new computer technologies as fast as drug companies can buy them.
Stephen McGarry, an analyst for Goldman Sachs, thinks there’s a good chance that the biotech and computer companies may eventually ally with one another, with computer giants taking a cut of drug royalties. If this comes to pass, perhaps bioinformatics experts by then will have succeeded at last in building computer systems that can discover new drugs. But right now, it’s not at all clear that they will.
