Gene sequencing may be cheap, powerful medicine

SAN JOSE, Calif. -- They'll be as small as your microwave and cheaper than your laptop. And the powerful tools -- future gene-sequencing machines -- will be able to tell you exactly what you're made of. Someday, they could help keep you healthy.

It took 10 years and $4 billion for the federal government to complete the just one sequence of the human genome in 2000. Its equipment filled vast rooms at many campuses.

Now rival scientific teams are racing to build tests that can accurately sequence an entire human genome in less than 30 minutes for $1,000 -- about a hundred-fold drop from current prices. The cheaper tests will make it possible to sequence the genomes of thousands more people, providing vital data about human traits, such as susceptibility to disease.

Virtuoso bioengineers predict that within two years, their tests will start to transform medicine much as PCs rocked the world of mainframe computing.

Modern genetics is essentially a giant math problem. Which of our 23,000 genes cause, or protect against, thousands of known illnesses? Do they work alone or in combinations? How are they affected by the environment?

As more genomes are sequenced, correlations between genes and diseases will become more apparent. Currently, scientists searching for rare genetic variants in a population are rooting for a needle in a haystack, said University of California at San Francisco Chancellor Susan Desmond-Hellman; the university just opened the Helen Diller Family Center Research Building, which will take advantage of the emerging technology.

As the cost of sequencing drops and it becomes practical to scan many more genomes, we'll get much better at "needle-spotting," she said.

Over time, sequencing will move into a more diagnostic settings, such as community hospitals. The tests could be used to decipher the genetics not just of healthy cells but out-of-control cancer cells. For instance, there are 38 different genetic subtypes of leukemia.

Genomic profiling could help doctors match the treatment to the precise disease, improving patient outcome and reducing side effects. That is because some cancer genotypes respond well to chemotherapy, while others are unaffected.

There remain formidable policy problems to solve, such as how to keep genetic information private. Also, there is a vast amount of information to be interpreted and stored.

The acceleration is being propelled by technological innovations. The performance of gene sequencers has accelerated even faster than that of microprocessors, which double every 18 months under the famed "Moore's law." Money is also a motivator. The first team to sequence 100 human genomes in 10 days at a cost of $10,000 or less per genome will be awarded $10 million by the X Prize Foundation. The market for full-genome sequencing and analysis could be worth billions.

The old approach used by the government's Human Genome Project worked by chopping DNA into short strands, sequencing each strand and then laboriously stitching them all back together.

Newer generations of the test, such as that offered by the Cambridge, Mass.-based Knome, are available to private individuals for about $100,000 and take about two weeks to complete.

"Technologies that are smaller and faster are driving this along," said Mark Stevenson of Applied Biosystems of Foster City, Calif. In 2010, sequencers will cost about $10,000, he said. The price tag will fall to $1,000 by 2011, as technologies such as nanosized chemicals and lasers become more perfected, he predicted.

The goal is a full-genome scan, far larger than the tests offered by existing gene-testing startups. These companies test only a tiny part of a person's DNA -- certain parts of a chromosome known as single-nucleotide polymorphisms, or SNPs. Variations in SNPs are linked to the likelihood of developing certain diseases. But they don't tell the whole story.

"Right now we have a very limited subset of information," said J. Craig Venter, a scientific leader in genome sequencing. "Soon we'll have not just one or 10 genomes but tens of thousands of genomes."