Slashdot Mirror

They didn't invent the wheel here, they are using a bead chip from Illumina. I imagine accuracy information is already readily available.

Some people are missing the point here, so for emphasis: this product only prepares DNA for sequencing, it doesn't do the sequencing itself. Half an hour of preparation is reduced to minutes, but the actual work still takes days.

It used to take days, and still does if funding is short but an Illumina HiSeq 2500 can produce 150-180 Gbases in 40 hours in rapid run mode [1]. Most labs still run it in high output mode because of the reagent cost but the option is there. This means that if I was prepared to pay the extra and I sent a sample into "core sequencing" where I work, they could potentially return mapped DNA in a week. After that there's still some improvement tools we'd need to run to clean up artefacts, followed by calling and filtering variants, those bits can take weeks. Whilst it is true that the bottleneck is currently the physical sequencing process of things but pretty soon that is going to shift to the informatics.side.

[1] http://www.illumina.com/systems/hiseq_2500_1500/performance_specifications.ilmn

Hey, you can always buy some of our stuff: http://www.illumina.com/technology/moleculo-technology.ilmn - 10kb fragments with Q50 accuracy should be enough for everyone!

In comparison, the sequence data people are producing today is crap. The individual reads are 30-80 base pairs and get put together into contiguous runs of only several thousand bases of length, on average. This is good for some kinds of work, but it doesn't give nearly the same picture of the genome that made the original human genome sequence such a masterpiece.

(I'm a genomics grad student. Can you tell?)

"Not at all. The issue here isn't statistics. It's that less-than-brilliant people could potentially be making huge, life-changing decisions based on this test."

Maybe Darwinism should take its course.

"Yes, the tests might give you an idea about your risks for certain health problems, but there is no real benefit to these kits. The same testing (or better) is available through labs, if someone feels it's necessary."

These genetic test kits are processed by CLIA-certified labs using state of the art equipment (Illumina for example - http://www.illumina.com/systems/iscan.ilmn ).

"Not to mention, as someone said above, family history is a pretty accurate (and free) way to get similar information."

Not if: 1) you don't know who your family is; and 2) there's a history of mis-diagnosis. Looking at a family that is now recognizing widespread Celiac Disease and gluten-intolerance, decades of past GI-related ailments in deceased family members including ulcers, gastroesophageal reflux disease, colon and esophageal cancer suggests a common basis that was never considered in individual cases. I've seen bad drug interactions diagnosed as epilepsy, peanut allergies diagnosed as candida to chronic infections, other food allergies diagnosed as Barrett's esophagus, allergies diagnosed as personality disorders, etc., etc., etc...

The kits that Walgreens proposed having in their stores seemed to have an element of bait and switch in them (low initial cost followed by a variety of add-ons), so the tests packages should be clear as to what the consumer is really getting. On the other hand there are big pharm companies making huge amounts of money off consumers who neither understand their meds, nor side-effects/contradictions, nor general effectiveness, nor how lifestyle changes can often be far more effective than the meds they're prescribed.

If we're going to have this, this "26 marker" stuff isn't good enough. Commercially available systems can do over a million markers per sample. "23andMe" uses those. There have been false matches with only 26 markers, but the modern systems that use tens of thousands shouldn't have that problem.

hmm..correct and incorrect, imho. Sure, there has been significant advances in the way sequencing works; the lastest being 454 sequencing http://www.454.com/, or Solexa http://www.illumina.com/pages.ilmn?ID=203 or SOLID http://www.illumina.com/pages.ilmn?ID=203, which has significantly reduced cost to sequencing. However, with each of these new techniques come new challenges in statistics and data analysis that are not just technological problems - they require significant, real breakthroughs in algorithms and statistical methods - how do you identify genes? what statistical methods would you use to identify distant repeats separated by millions of years? How accurate and reliable are these identification methods? We've come to a point where getting the data is now almost trivial and cheap - making sense of it, even being able to order it in the right way - we're just beginning to make headway there. So its not all tech - there is a lot of science there - only, it is difficult to argue it is biology any more - more chemistry and math and a bit of CS :)

hmm..correct and incorrect, imho. Sure, there has been significant advances in the way sequencing works; the lastest being 454 sequencing http://www.454.com/, or Solexa http://www.illumina.com/pages.ilmn?ID=203 or SOLID http://www.illumina.com/pages.ilmn?ID=203, which has significantly reduced cost to sequencing. However, with each of these new techniques come new challenges in statistics and data analysis that are not just technological problems - they require significant, real breakthroughs in algorithms and statistical methods - how do you identify genes? what statistical methods would you use to identify distant repeats separated by millions of years? How accurate and reliable are these identification methods? We've come to a point where getting the data is now almost trivial and cheap - making sense of it, even being able to order it in the right way - we're just beginning to make headway there. So its not all tech - there is a lot of science there - only, it is difficult to argue it is biology any more - more chemistry and math and a bit of CS :)