Are Genetic Tests Useful Part 2

PREV 3 of 3

For starters, none of the sites tested me for the rs10757278 marker, despite its importance as a predictive marker in several studies. The reason is one of expediency.

It turns out that this marker does not appear on the gene chip used to analyze DNA for the DeCodeMe website. This chip—called an "array"—is made by San Diego-based Illumina, which sells "off the rack" arrays used by deCodeme and 23andMe, and includes only about a tenth of the 10 million genetic markers that scientists believe account for much of the diversity in humans.

For my original test run by DeCode, they used a custom-designed array that tested for rs10757278. For their consumer site, they used the standard chip, not their custom one.

But the Illumina array does test for other DNA markers that are closely associated with rs10757278. They all occur in the same region on chromosome nine, and each confers a similar heart attack risk factor (a linkage disequilibrium, for all you scientists out there). One of these linked markers is rs10116277—which deCodeme used instead of rs10757278.

When I called DeCode to ask about the contradiction between my results from the custom chip and from the consumer website, company scientists said that I was an anomaly. Most people, they said, test the same for both of these markers. That is, if they are at a high risk for rs10116277, they test the same for rs10757278.

I'm part of a minority that tested differently for the two markers, with one coming out as high risk (red), and the other average (black). DeCodeMe also tested for a second genetic marker on chromosome 1 that gave me a below-average risk.

Navigenics tested me on a second marker linked to rs10757278—they use a different off-the-rack array than deCodeme—and I came out high risk again. This is why their score for me is in the upper range of danger.

Michele Cargill, head of Translational Genetics for Navigenics, told me that their marker is a perfect linkage disequilibrium for rs10757278, though the studies cited on the Navigenics site do point out that for a small group of people, these markers are not identical.

"We're just beginning to understand what this information looks like for individuals," acknowledged Cargill, "but we think there is enough information to tell you things that will help you make decisions about your lifestyle and health."

Heart disease isn't the only sticking point. The problem of conflicting results—based on looking at different genetic markers for the same disease—is also apparent in my results for type 2 diabetes, rheumatoid arthritis, age-related macular degeneration, and more. It all leaves this consumer a bit confused about what to believe.

As these tests evolve, scientists will need to sort out why some people have contradictory results for individual gene markers. Standards also need to be established for which markers are most relevant, and how risk factors are determined for single markers and groups of markers.

As geneticist Francis Collins, director of the National Human Genome Research Institute, said Wednesday during a meeting I attended in Washington, D.C.: "We don't want to squash these companies, but there needs to be some oversight. We need a public site to show consumers what is accurate and valid. Otherwise we risk turning off the public."

This will come, but, at the moment, this consumer can only suggest that buyers beware the pitfalls in these early days of this promising technology.

Next Week: Genes for wet or dry earwax and other results and ruminations about recreational genetics and genetic voyeurism (the privacy issue).

PREV 3 of 3