A couple of weeks ago I reported on a presentation by 23andMe's Nick Eriksson at the American Society of Human Genetics meeting in Honolulu, in which Eriksson presented data on a series of genome-wide association studies performed by the company using genetic and trait data from its customers.
Along with genetic analysis of a variety of other traits (such as asparagus anosmia and photic sneeze) Eriksson presented data on two novel regions significantly associated with hair curl, one close to the TCHH gene and a second near WNT10A (see the abstract for details). I noted at the time that 23andMe appears to be doing a pretty good job of running genome-wide association studies, although of course the real test of this is independent replication.
Well, now we have replication (of a sort) for at least two of 23andMe's novel findings - but unfortunately for the 23andMe crew the "replication" study has beaten them into print.
In this week's issue of American Journal of Human Genetics an Australian group report a genome-wide association study of hair curl that uncovered both of the novel regions picked out by 23andMe's scan (the very strong TCHH association is the headline finding, while WNT10A is mentioned briefly later in the text).
In a sense this is good news for 23andMe, in that it's good, independent evidence that the company's novel associations are robust. However, it's also bad news: not only has the company been scooped on the hair curl findings, but this is a clear illustration that there are few phenotypic niches sufficiently obscure for 23andMe to be able to work safely without worrying about competition from academic consortia.
The Australian group that produced the hair curl study (led by Nick Martin) has both the resources and the expertise to churn through the same sorts of commonly variable non-disease traits that 23andMe is well-powered to analyse, as their publication list illustrates. Other rivals have even greater capacity for scoopage: ailing biotech deCODE, for instance, has genetic and detailed trait data for a hefty chunk of the Icelandic population that can easily be used to unravel the genetic architecture of whatever common trait it decides to turn its attention to.
One major advantage that 23andMe possesses over any of its rivals is a highly actively involved participant base, many of whom will remain available and willing to answer online surveys and purchase additional genetic tests so long as the company can keep drip-feeding them interesting nuggets of genetic information about themselves.
Is that enough to maintain a profitable business? The two waves of lay-offs (one in June, and a more recent one in October) and the recent departure of co-founder Linda Avey from the company suggest that there may be some uncertainty about that even within 23andMe. Still, there's little question that the company has done more to establish the personal genomics industry than anyone else, and it's still far better-placed than either of its two major rivals, deCODEme and Navigenics. There are big changes coming in personal genomics, but I don't expect to see 23andMe toppled from its pedestal any time in the near future.
Life Science
Medicine & Health
Comments
Is it not expected to be 'scooped' at some time especially in such a competitive field? Methinks, 23andMe are getting a taste of what it's like for all other researchers out there!
Posted by: mell | November 11, 2009 7:36 PM
Has anybody out there actually purchased any of the DTC testing? What is the experience like? What do you learn? How much does it cost? Do you have to take the info to a genetics counselor or physician trained in genetics or anybody who can actually interpret these data? And then what? Interested in hearing from anybody who has personal experience with this stuff. Do you really *spit* into something? (Yuck!)
Posted by: Catharine | November 11, 2009 8:17 PM
Daniel,
What are the big changes in personal genomics that are coming?
-Steve
Posted by: Steven Murphy MD | November 11, 2009 8:58 PM
@Catherine I just got my report from 23andme a couple of weeks ago. It's a bit expensive (normally, $400 though they were having a sale when I got mine). They have very nice reports and explanations, no special training is required to understand the results. Their site will tell you what things they report on. For me, I was pleasantly reassured that I haven't picked up a tendency towards a couple of late-onset minor ailments which run in my family. Then again, they also say my eyes are most likely to be blue (they are green) so that helps remind me of the limits to the current state of the art.
Posted by: Nox | November 11, 2009 11:07 PM
alternate views.
1. Science: Look how quickly this has been published and reconfirmed.
2. Culture: Small independent research teams can still 'beat' large commercially funded research.
It is behind a paywall, Did they confirm the both specific rs17646946(A;A) and rs7349332(T;T) reported by 23andMe, or others in that gene?
Posted by: cariaso | November 12, 2009 3:28 AM
Well, I wouldn't call Nick Martin's group "small"! Still, it's definitely easy to spin this as a triumph of academic curiosity over commercial research.
As for the SNPs: rs17646946 is the most significant SNP genome-wide (P=1.5x10^-31) and rs7349332 is the most significant SNP in the WNT10A region and the 25th most significant genome-wide (P=1.4x10^-6). So yeah, they've hit the same SNPs, not just the same regions.
Posted by: Daniel MacArthur | November 12, 2009 4:42 AM
@cariaso
Nick Martin's group is a very large research group, with a huge biobank of phenotyped samples, and an army of researchers, students and support staff. 23andMe is less than half the size in total, with the 23andWe research being carried out by a handful of researchers.
If anything, this is an example of the unsurpising result that a very large, well funded research group can beat a small, cash-strapped company to get research published first. "The underdog lost" is never a particularly interesting story.
As it happens, they both got the results at about the same time, and Nick Martin's grouped published slightly faster; I don't think we can really draw any conclusions about public vs private research here.
Posted by: Luke | November 12, 2009 6:25 AM
The whole notion of being "scooped" highlights a huge part of what's wrong with the current research model...that it's a competition of some sort. This is why people hoard their data and don't collaborate openly with others. Patenting is partly to blame, as well as the all-consuming quest to get that next big grant and, oooh, maybe even the Nobel prize (!). Research shouldn't be about the researchers..it should be about the patients.
Posted by: Gene | November 12, 2009 7:15 AM
Hi Luke,
Agreed. I understand the 23andMe paper was actually submitted quite a long time ago, so in fact the most on-target narrative here probably relates to the disturbingly arbitrary nature of the peer review process rather than any simple "us vs them" story.
Posted by: Daniel MacArthur | November 12, 2009 7:15 AM
Or was it the fact that the research was done on Human Subjects with no IRB process???? Maybe that's why the 23andSerge paper was held up???
Thoughts?
-Steve
Posted by: Steven Murphy MD | November 12, 2009 8:18 PM
is there a way to just 'ignore' or filter out the trolling comments from Steve, Daniel maybe you can add that as a feature to your blog :)
Posted by: Ricardo | November 13, 2009 10:29 AM
@Catharine
"Has anybody out there actually purchased any of the DTC testing? What is the experience like?"
I purchased a 23andMe v1 test ($1000) and the v2 upgrade ($400). I've also ordered for myself Myriad BRCA testing (3 SNP $400), and I've done some hobby PCR testing on myself.
Genetic testing yourself is like running diagnostics on your home computer. Unless you know what you want to learn, you're probably not going to learn anything useful, no matter how excellent the report.
I run Steve's medical practice. Our problem in practice is not data itself, but sorting and applying data for some use. Anybody can read the results of any medical test, just like how anybody can read the law or anybody can read computer code. The challenge is to understand all the information about a patient to achieve something useful when the patients themselves have no literacy nor ability to communicate information about themselves. It's a problem of significant digits: (10 minutes of attention) * (two vague phone messages with a secretary) * (decade of loose hand-written notes) ...
Genetic testing is immensely useful. But, so are computers, and most people use computers for complaining to their children about how computers don't work. Or, "Magic Nigger Fight: ONLINE!" Too bad we can recreate an entire virtual Earth to pretend to slaughter neighboring villages for shiny loot and boobs, but we can't even agree on how to send a patient's name in an electronic medical record.
"Would you tell me, please, which way I ought to go from here?" "That depends a good deal on where you want to get to," said the Cat. "I don't much care where--" said Alice. "Then it doesn't matter which way you go," said the Cat. "--so long as I get SOMEWHERE," Alice added as an explanation. "Oh, you're sure to do that," said the Cat, "if you only walk long enough."
Posted by: Andrew Yates | November 13, 2009 1:47 PM
@Ricardo
"is there a way to just 'ignore' or filter out the trolling comments from Steve, Daniel maybe you can add that as a feature to your blog :)"
Let me filter Steve's excellent point from "Incredibly Intelligent Medical Doctor Who Does Not Make His Patients Feel Stupid So That They Resent Him Despite Telling Them That The Way They Live Their Lives Is Wrong" into "Polite Passive-Aggressive Nerd Comment Forumese" for you.
"Or was it the fact that the research was done on Human Subjects with no IRB process???? Maybe that's why the 23andSerge paper was held up???
Thoughts?
-Steve"
>__
Published medical research has a form to expedite review including academic affiliations and institutional review boards. Deviations from this form demand more attention from reviewers and publishers, and this increased load may be sufficient to retard publishing compared to identical research without the deviations in form.
;)
Posted by: Andrew Yates | November 13, 2009 1:59 PM
hmm, you don't by any chance work for Steve do you?
Posted by: Ricardo | November 13, 2009 11:01 PM
@Ricardo
Yes, Steve and I work together in the same office. I am not employed by Steve.
from #12 "I run Steve's medical practice."
Steve has an obnoxious writing style, but he connects well with patients because he doesn't make them feel stupid or act aloof.
I didn't like 23andMe 6 months ago, but I like that they seem to be moving in a new direction. I don't know what that direction is yet, but I'm optimistic.
Posted by: Andrew Yates | November 14, 2009 1:44 PM
Dr. Yates, I could never understand why someone of your obvious intelligence is working with Murphy. He may be the picture of wit in person, but his public persona defines him. If you have any influence over the man, could you tell him to (1) write in complete sentences, (2) stop drinking and blogging for the children's sake, and (3) learn that an ellipsis involves three and only three periods? The internets thank you.
As for the topic at hand, in light of:
1) the firing of 10 employees in July
2) this brutal review by an employee on glassdoor
3) the firing of Linda Avey
4) and the firing of 18 more employees just a few days ago
...does this mean that you are rethinking your comments from about a year ago on "23andMe == DTC Genomics now and forever"?
In your defense, there is a new fact in the offing, whose importance over and above business model you did apprehend at the time. It seems that either Sergey Brin or his wife have voluntarily decided not to pour (more of) his personal billions into the enterprise, perhaps because the mothership of Google is now running into squalls and it would look unseemly for them to eat nepotistic cake rather than saving his 'real' employees (to mix metaphors).
Posted by: asdf | November 15, 2009 8:59 AM
@asdf and @Ricardo
sometimes my mouth moves slower than my mind and even slower than my fingers.
But the message is clear, this phase of empowerment in genomics is a blip and no one is prepared to change for a blip. Whether it be journals, physicians, scientists or any patient or rational person.
Despite the hype, the reality is crystal clear; other than a band of fools willing to sell their genome to google for some bobbles (web portal) most people are unwilling to take that leap, thus the market has already been sold.
Steve
Posted by: Steven Murphy MD | November 16, 2009 7:21 PM
@asdf
...does this mean that you are rethinking your comments from about a year ago on "23andMe == DTC Genomics now and forever"?
Oh, I think the next wave of genetics will not be called "DTC Genomics." So, no, I think 23andMe will stand alone in that category so long as anybody cares to remember it by that name.
And, hey, Bush is an idiot, but that idiot was president for 8 years.
Um... I don't know how that defends Steve. But, it's true!
Posted by: Andrew Yates | November 17, 2009 4:54 PM