Opportunities Abound in Pharmacogenomics

if (n == null) The Scientist 13[10]:16, May. 10, 1999 Profession Opportunities Abound in Pharmacogenomics By James Kling Graphic: Leza Berardone "Have it your way!" proclaim the restaurant chain's ads. What is arguably true for patrons of the fast-food industry is likely inevitable for patients seeking therapy for their illnesses. Pharmacogenomics--the application of genotyping to patient therapy--could revolutionize drug discovery and medical treatment. Side effects plague most prescription drugs--often, some percentage of patients gain no benefit at all. Many believe that this can be traced to genetic differences that disrupt the interaction between the drug and its target, or that show up in different stages of a disease, as in many cancers. With the advent of the Human Genome Project and automated analysis that can handle up to 25,000 samples per day, researchers predict that in the near future scientists will understand the vagaries of genetic pathways of major diseases well enough to custom-fit therapy based on a genetic profile. Physicians will run a battery of genetic tests on the patient before sending her off to the pharmacy for the drug that is the odds-on favorite to cure what ails her. The promise, however, of pharmacogenomics is a long way from fulfillment. Companies will continue to develop assays, build new and better-automated instruments, and design and carry out the epidemiological and genetic studies that could form the backbone of 21st-century medicine. That boon will create a plethora of jobs and career opportunities for life scientists. "There is every reason to believe that pharmacogenomics is an area with enormous growth potential--it will become a major theme in pharmaceutical [research and development] and medical practice," says Charles Cantor, chief scientific officer of San Diego-based Sequenom Inc. Mining for Information Naturally, computer skills will be in high demand. "A molecular biologist with computer skills will be immensely useful. I don't think we're that far away from labs with no wet chemistry being performed at all--just scientists in front of computers," says Chris Ashton, marketing director at Oxford GlycoSciences PLC, Oxford, U.K. (See related story, "The Virtual Cell") That trend has already begun with the informational tools--such as the National Institutes of Health's genetics database GenBank and protein sequence databases--that are available to biologists. One of the best ways to prepare for pharmacogenomics, advises Ashton, is to streamline your research using those tools. "If someone has already done an experiment and it's documented, why [repeat] it? It's increasingly difficult to find someone who really knows how to mine for information using computers." Russ Granzow, director of business development at Orchid Biocomputer Inc., Princeton, N.J., sees a need for life scientists with an interest in improving search engines and data-mining tools. He believes researchers will conduct upward of four billion genotypes over the next three to five years. The technology is already in place to churn out those samples, he says, but the vast dataset that will result from it will pose a huge challenge to scientists hoping to analyze it. The dataset includes four billion genotypes, about one million genetic loci with medical relevance, anywhere from 400 to 1,000 compounds currently on the market, 100,000 human genes, perhaps 400 diseases of commercial importance, and 100-200 million patient medical records. "One human brain can't handle all of that information," comments Michael Kauffman, vice president of medicine at Millennium Predictive Medicine Inc., Cambridge, Mass. Kauffman cites a yeast study his company is conducting. Specific drugs inhibit the synthesis of the yeast equivalent to cholesterol. To better understand the mechanism of these anticholesterol agents, and to design new drugs that can circumvent resistance, researchers measured the expression of RNA levels after exposure to a particular drug. Programmers then designed software to organize the data and present it to researchers in an intelligible form. They began by grouping changes by time frame--assuming, for example, that RNA changes that occurred in the first hour after drug exposure were probably related to each other in the biochemical pathway. Then they grouped RNAs with similar base sequences. They also linked the RNA sequences to a biochemical pathway database. The linked data ultimately led to a three-color printout: One color represented genes that dictate the biosynthesis of yeast's cholesterol equivalent, another set included genes involved in the cellular stress response, and the third represented genes of unknown function. "Now you have biologically meaningful information, and you can then go to a yeast strain that is resistant to the drug. What genes are expressed in the resistant yeast? When you do the same kind of analysis, you find that different genes are turned on by the drug in a resistant strain. So you build the next generation of drugs directed at the resistance genes," explains Kauffman. Professionals will be in high demand to produce similar systems that analyze massive databases linking adverse reactions to particular drugs with genetic profiles. They'll also be needed to tease apart the biological mechanisms that cause different clinical manifestations of cancer, Alzheimer's disease, and even highly variant infectious diseases such as AIDS. Polydisciplinary Scope It may be a brave new world, but companies don't discount the value of a meat-and-potatoes scientific education. "It's hard to go to graduate school in genomics, and I don't recommend that you do. Get the most rigorous molecular biology training possible, and include biochemistry. There has to be a focus on biological pathways, and some integration with disease. We're all in the business of fixing diseases," remarks Kauffman. "Epidemiologists and classically trained statisticians do have an opportunity to play a role," notes Granzow. "Informatics is computer programming, but it's also accessing information from much broader fields" of study. "We are a totally polydisciplinary company," says Cantor. "We have surface chemists, nucleotide chemists, people who can design and build instruments ..., and a lot of molecular biologists who are interested in developing novel assays. We also need people who can help us work with epidemiologists--most of the people we collaborate with are large entities with their own epidemiology groups. But we need people here who are counterparts--otherwise we can't talk to them on an even playing field." Molecular Technology For More Information GenBank www.ncbi.nlm.nih.gov/Web/Search/index.html Millennium Predictive Medicine (Millennium Pharmaceuticals) www.mlnm.com Orchid Biocomputer www.orchidbio.com www.snps.com Oxford GlycoSciences www.ogs.com Sequenom www.sequenom-san.com The Automation Partnership www.autoprt.co.uk There's plenty of work for software aficionados, but the new techniques and applications will also put greater demand on hardware development. Some liken their mission to "building a factory," where high-throughput samplers and analyzers are the rule of the day. Those instruments are built and maintained by technicians trained in molecular biology, analytical chemistry, biochemistry, and other fields. The technicians "have a penchant for technology--they're molecular technologists," says Granzow. Those scientists work with engineers, optical physicists, and analytical chemists on surface immobilization of DNA samples, getting enzymes to function in minute volumes of microreactors, calibrating the optimal amount of electric current to keep polymerases active in small volumes, and converting fluorimetric or calorimetric measurements to other forms of detection. Molecular technologists can also expect to get friendly with mechanical engineers. Wires and chips aren't enough to design instruments that can fit biological research needs. "You've got to have the sort of scientists who have a deep understanding of the biology, who can work with mechanical and other engineers to produce a product," comments Ali Ahmadi, product manager for genomics at The Automation Partnership Ltd., Cambridge, U.K. An Active Industry The intense competition between rival technologies and the heightened interest in pharmacogenomics make the industry a challenging career choice. The industry places a premium on carefully planned, insightful experimentation, says Kauffman. "Say you're doing 25,000 analyses (using an autoanalyzer) every time you do an experiment--if you do the wrong experiment, you waste a lot of time. Our best people are the ones who know how to set up the right experiment. People who don't do so well set up the wrong experiment or they make a mistake setting it up and spend a long time analyzing irrelevant data." Pharmacogenomics laboratories generate an intense working environment. "We want people with a ton of energy. You can almost tell when they walk in [for an interview] if they have that energy level. If you don't have it, you just think [we all] are nuts--and I won't disagree with that," asserts Kauffman. Unlike most of the biotechnology world, pharmacogenomics promises a much more immediate impact on patients, explains Granzow. Pure drug discovery is 10 or 15 years away from any practical benefit, and even research that provides drug discovery tools or diagnostics is still a few years away. "Pharmacogenomics is unusual because it's in the last phase of drug discovery, and it's being applied to drugs that are on the market. [Patients and doctors are] out there clamoring for this stuff--so that's a very fulfilling thing." James Kling is a science and technology writer in Bellingham, Wash. His Web page can be found at nasw.org/users/jkling.

Opportunities Abound in Pharmacogenomics

Opportunities Abound in Pharmacogenomics

By James Kling

Interested in reading more?

Become a Member of