Emily Singer has been doing a great job of covering the consumer genetics beat over at Technology Review; her most recent piece draws on a recent presentation by former head of the National Human Genome Research
Institute Francis Collins.
Collins caused a bit of a stir during his presentation (at last week's Consumer Genetics Show in Boston) by announcing that he had signed up for several personal genomics services under a false name. His conclusions:
Collins said that sequence-wise, the tests "appear to be
highly accurate": there were almost no differences in the genotype information generated
in the three different analyses. But there were significant differences in the
numbers of genetic variations used to calculate disease risk, as well as the final
risk score. For example, one company used 5 single nucleotide polymorphisms,
or SNPs, to calculate risk for a particular disease, pronouncing Collins at low
risk. Another used 10 SNPs, placing him at high risk, and the third used 15, concluding
that he is at average risk. Collins also said that the analyses provided little information
on his "carrier status," meaning whether he carried genetic risk factors that didn't
influence his own risk of disease but could be passed down to future
generations.
There's nothing particularly new here, except that these words come out the mouth of a science rock god. For instance, the high technical accuracy of these tests is unsurprising given that personal genomics companies rely on the same technology developed and painstakingly refined for use in massive genome-wide association studies.
The differences in risk prediction algorithms between companies is a chronic concern - the companies can even differ substantially in their estimates of the background risk of common diseases - that makes it difficult to compare results; there have been noises made about fixing these discrepancies for some time, but they persist. Finally, personal genomics companies (especially 23andMe and newcomer Pathway Genomics) currently provide some limited carrier testing, and stealthy outfit Counsyl appear to be aiming to provide a much more systematic assessment of carrier status.
Thus Collins' concerns about the performance of personal genomics companies seem to be technical and eminently soluble, rather than representing any deep ethical problems. That bodes well for personal genomics companies given Collins' high status within the ranks of science policy advisors to the current US administration (the rumours that he will become the next head of the National Institutes of Health are pervasive enough to have become boring).
Moving beyond consumer genomics, Collins's assessment of the impact of recent genome-wide association studies (GWAS) is optimistic - perhaps overly so:
"We have undervalued these studies," he said at the
Consumer Genetics Conference. "Even if a variant has a small impact on disease risk, that doesn't
mean it's not a good risk target." In type 2 diabetes, for example, two of the
nine common genetic variants that have so far been linked to the disease are
involved in the pathway targeted by two major diabetes drugs. "[Pharma
companies and others] have not jumped on this as rigorously as they could,"
said Collins. "Perhaps because it's a bit overwhelming--there are so many of
them."
GWAS signals certainly have the potential to teach us much about the underlying biology of common diseases, but it's important not to understate the barriers between finding association signals and designing effective therapeutics - especially given that we still don't understand the biological origins of the vast majority of GWAS hits.
It's good to see a likely future NIH head with such a bright vision for genomics, but I hope Collins' doesn't end up over-hyping the immediacy of the impact of genetics on healthcare. Such a move risks burning out the public before the real benefits of personalised medicine start to become readily apparent.
- Log in to post comments
I'm concerned that Collins doesn't think that epigenetics doesn't represent a far more plausible area for assessing the heritability of disease risk. While GWAS has had its successes, it seems that the vast uncharted areas of disease susceptibility due to methylation, histone protein markers (see recent advance publication in Nature) and other epigenetic information should be a major driver of future studies rather than dumping more money into GWAS.
As an illustration I'm thinking of GWAS = squeezing the last drops out of a lemon, epigenetics = crate full of lemons full of unsqueezed juice.
On the plus side it seems that the funding is moving in this direction as the overall message re:GWAS has been "not enough benefit to justify additional money/effort".
Actually, Collins is a huge proponent of epigenomics and initiated the large human epigenome project shortly before he stepped down as head of NHGRI.
Can you specify the specific Nature paper you are referring to? I can't seem to find it.
Francis Collins once said "the worst thing we can do for Personalized Medicine is oversell its promise...."
I don't think his is inclined to hype.....too much.
-Steve
Ok, this discussion [cf anon #1's post] is getting a little inside the beltway, with folks lobbying for more support for their favorite biomarker. Before weighing in with my $.02--and to avoid getting into a lose-lose p***ing contest about whose subfield is more deserving of funds--let me emphasize that we (as a research community) should be trying many different approaches. GWAS have their strengths--including relative inexpensiveness and mature technology, design, and analysis--but they also have a lot of well-known limitations. Epigenetics is clearly important biologically, but presents important design challenges--eg getting the right tissue at the right time--and (from the perspective of a distant onlooker) there seem to be a lot of technological bugs to work out to get assays that measure specific methylation sites on genome-wide level quickly, cheaply and accurately. Ditto sequencing, the technical and analytic challenges of which Genetic Future has covered admirably well. That's not to say we shouldn't invest in solving those bugs--we definitely should and are [cf anon #2's comment]--just we shouldn't abandon the proven GWAS approach when it still has much to tell us (including providing candidates for epigenetic and sequencing studies).
My read of the situation--and it is just that, my subjective take, based on the little piece of the sky I can see from the bottom of the well where I live--is the winds are blowing against further funding for GWAS, as the perception is they have nothing more to teach us and what they have taught us is not interesting. I think that perception is wrong on both counts, so I hope my worries are misplaced.
Hi Peter,
Excellent points on the challenges of pursuing epigenetic risk, although we all seem to agree that it will be important to eventually overcome these challenges.
Interesting that you see the winds shifting against GWAS - so you think that the next round of chips (adding in the low-frequency variants from 1000 Genomes) won't be enough to pique the interest of funding bodies? I suspect there will be some fairly impressive sales pitches from the chip-makers, especially the increasingly desperate Affymetrix, as well as from advocates of the rare variant hypothesis. It will be interesting to see if that's enough to hold back a full-pace sprint into sequencing.
Nature paper
Re: Peter's comment, I agree that some of the technologies for quantitating methylation are evolving quickly, and significant funding is needed there to push some of the technologies forward and get them hammered down so that additional discoveries can be made for patients with disease that have been let down by traditional genetic approaches.
My main point is that I think there is far more potential wins in this alternate approach, especially considering how we know that epigenetic profiles can be changed with lifestyle choices and drug interventions. If we better understand how the evolving epigenetic map interacts with the genetic landscape, this would represent a huge new target for improving our ability to diagnose and treat disease.
Hey anon,
Just a heads-up - at some stage in the near future I'll be implementing a policy blocking commenters from using "anon", "anonymous" or related handles as it can lead to confusion between commenters. It would be better if you could pick another (similarly anonymous but more unique) handle.
Sorry to distract from the main thread, but thought I should let you know.
Daniel,
Thanks for the heads up and for the great work you do here. I'm a big fan.
bb
I'm with Peter too. There's a lot more mileage in GWAS than we've plumbed to date - which is not to distract from other modes of inquiry. I think most people appreciate that no single M.O. is going to prove sufficient to answer any question (in this case: what are the genetic determinants of disease risk/quantitative traits).
Might be the wrong place to post this...but...my general worry about all the massive data approaches to date is the reification of the gene as a basis for disease. The answer to the missing heritability problem is unlikely to be an epigenomic answer. I know there are strong political motivations pushing systems biology...but at the end of the day "where's the beef" folks. Pushing more money to the largest U.S. universities has certainly concentrated money in "centers" of dubious excellence or relevance...but the long term value of the data does not seem to be debated enough. It worries me that traditional Genetics, molecular biology and developmental genetics isn't getting stronger support, and I worry that whole sectors of the U.S. research enterprise (small and medium size universities) are being frozen out by a political shift of funds to unproven tech...so..its 20 years on...where's the beef?
:-)
Oh..and speaking of pharma...yah they are a little shy about investing $100 X 10^6 in a small molecule aimed at a polymorphism that contributes (maybe) 3% to risk. Duh Francis. Even the "network" folks have brought barely anything to stage II trials...and the efficacy data is driving folks back to running non-profits...
so..donde es el carne de vaca? :-)..not too trolly I hope but...I like a little science along with my marketing.