New poll: would you make your genome public?
Would you have your genome sequenced for free? Conditions: you must license it for all use; a liberal CC-no attribution-like license which allows for commercial use as well. Also, your genome will be made public with many personal data such as age, height, sex, weight, ethnicity, personal status (we want to find the “money making gene” and the “fecundity gene”) medical history, some family history but not name or exact geographic location. Note that these metadata may make you traceable, although not easily. To fully appreciate the genotype-phenotype connection, your doctor will upload your medical records with each visit & checkup. Hey, once you contract that late-onset Cheetos addiction at the age of 65, we want to know about it. Hmmm, not so free after all, maybe. Still you may want to do it. Tick in your answer in the blue box in the right sidebar. HT to Mickey for the idea.
It would have been useful to give the option of making a genome public without any associated meta-data – this is obviously less useful for research, but still worthwhile.
No way – MotherJones did a great piece on this type of phenomena entitled “Google’s Guinea Pigs” November/December 2009. It was all about 23&Me. It’s actually pretty scary how much cash is being poured into this effort (sequencing individuals and linking their sequence to their online medical data).
http://motherjones.com/environment/2009/11/googles-guinea-pigs
The Mother Jones piece focuses on the current immaturity of our knowledge linking SNPs to traits and on the potential for overinterpretation and unscientific opportunism — in short, all of the scare-mongering and none of the upside.
Not exactly a balanced article.
23&me do SNPs, not genomes. Also, they do not provide that information publicly. So not a good example.
That article in Mother Jones is terrible. It is hard to go through all the inaccuracies, but here are the lowlights:
1) Brownlee has a very poor grasp on statistics, as evidenced by her cherry-picking outlying data in her SNP-phenotype correlation, and showcasing them as evidence of wholesale inaccuracy.
2) Her Warfarin example is false. There are guidelines on changing Warfarin dosage given the CYP2C9 and VKORC1 variant data. http://www.medscape.com/viewarticle/561608 These guidelines are not yet FDA directives, because the clinical trials are still underway:
http://www.genomeweb.com/dxpgx/wash-u-partners-kick-five-year-warfarin-pgx-dosing-trial
3) The prostate cancer example confuses early detection with genetic susceptibility. It may or may not be true that early detection does not help (Brownlee does not cite the research articles upon which she bases her opinion), but early detection of cancer already in progress is not genetic susceptibility. Would Brownlee advise women not to take the BRCA1 variant test, which has saved many lives? Would it not be the patient’s choice whether she should take the test,with the possible ensuing ordeal of mastectomy and re-constructive surgery?
4) Lumping 23&me with snake oil companies that do genetic matchmaking is, frankly, ridiculous. While 23&me is in the field of “recreational genomics”, they do ground their predictions in research, despite Brownlee’s poor grasp on the kind of data this type of research provides. Snake-oil remedies have always been around. Every new medical advance has brought a concurrent “snake oil” charm with it, personal genetics is no exception. Does personal genetics and the ensuing personalized medicine offer an opportunity for fraud? Absolutely. Should we therefore forego it? Or maybe subject it to the same stringent rules all medical fields are subjected to.
5) In her interview with Linda Avey, it is not clear who the “we” Linda refers to are. At the time of the interview, Avey has already left 23&me, and has founded her own Alzheimer’s research foundation, something Brownlee neglected to mention. In any case, the whole GoogleHealth+23&me conspiracy theory she hints at falls apart at the seams: nobody is forcing you to merge the two, nor are you forced to open a Google Health account in the first place (or spit for 23&me for that matter).
6) Brownlee seems to think that knowing of genetic predisposition to diseases may lead to unnecessary anxiety in the case of conditions that cannot be changed (she cites Parkinson) or conversely, runaway health-consumerism for those that may be changed. Her cure: a blanket moratorium on all genetic testings. Frankly, that’s asinine. Willful ignorance may be her choice, and I respect that. It would probably be mine for a SNP correlated with an incurable, early-onset disease. But some people may want to know, to settle their affairs in time, maybe retire early to enjoy life more, etc. Why deny those people that knowledge? As for runaway health consumerism: that would only be the privilege of the rich, who are not bound by health insurance regulations, which would probably be strict enough not to include treatments for low-correlation SNPs. After all, they are doing much worse in not providing treatment for existing ones…
@DM & @Iddo
I think you have both missed the point of the MJ article…entirely.
The article is about a journalist who decided to try the 23&Me product and then write about her experience. This person is not a bioinformaticist, nor a molecular biologist, nor a geneticist (to my knowledge).
This is exactly the type of person at whom this product is targeted. SNPs maybe low correlation, but that doesn’t stop A LOT of bona-fide scientists from publishing the work (peer-reviewed). In fact, for a long time (and even now) SNPs are used for disease correlation (aside: a lot of these papers are in PLoS and its spawn). I think that this type of testing is probably more popular than some in the scientific community realise ( http://www.thegeneticgenealogist.com/2010/01/07/personalized-genomics-a-very-personal-post/ ), and the fact that 23&Me was featured on Oprah probably lent it “credibility” with the masses.
Even with all the problems with SNPs, no doubt insurance companies are, or would be, willing to accept any type of predictor with the appropriate endorsements. I think the better question is how is 23&Me able to offer such a service for $500? That price seems low, even if they are fully automated (robotic processing) and using “established” primer sets.
I don’t really know what the Linda Avey having left 23&Me has to do with the information regarding 23&Me wishing to pair Google Health accounts (medical histories) with SNPs data. It seems quite likely that if (23&Me + GHA) can be shown to be “predictors” (whatever that means) of certain aliments, I’m sure someone will buy the data. And if people are willing to accept the 23&Me data, it isn’t too unlikely that they’d be amenable to obtaining a Google Health Account.
I wouldn’t want to have any genetic testing done myself. And I don’t think that’s ignorance. I think you are much more than a predisposition to X, Y & Z. Besides, what happens if you find out about a horrible disease? The knowledge of that could give you enough stress alone that could actually depress your immune system thus making you susceptible to the disease.
– CH
It’s really amazing to see all of the hype about the mass sequencing of thousands of human genomes, even if it will soon only cost a thousand dollars or less for each. While we are all about 99.6% genetically identical, there can be profound consequences of SNP variations amongst individuals in limited cases. Nevertheless, at the protein level there is an incredible amount of feedback and redundancies to compensate for variations in protein performance as a result of these SNP changes.
Genotype does not equal phenotype as the environment plays an equally critical role in the development of most known diseases. Less than 2.5% of the human genome appears to encode about 23,000 different proteins, which appear in perhaps over 100,000 alternative splicing forms and are subject to extensive post-translational modifications. For example, there appears to be as many as 500,000 phosphorylation sites in human proteins, which act as on/off switches to regulate the activities, interactions and degradation of proteins. Genome sequencing does not provide this information. The complex interplay of proteins is only hinted with genomic analyses.
It is ludicrous to think that we can correlate complex human behaviours and disease manifestations primarily with genomic data. At best, such data might have some utility for prediction of the health risks of large populations, but it will be fairly useless for individual diagnoses and prognoses except for a very small number of cases. Moreover, such information may not be very helpful to individuals if there are no effective treatments available. During complex embryonic development, the most serious of genetic defects are identified and these pregnancies are naturally terminated early on. Decades of epidemiology studies without the benefit of genomics analyses have already uncovered critical environmental factors to promote wellness and reduce the risk of disease.
A deluge of genomic data is drowning researchers and to a large part is already ignored by the vast majority of investigators. It seems that the mass gene sequencing efforts will only serve to make the hay stack of data magnitudes larger so it will become even more difficult to weed through and find useful discoveries rather than false leads. More effort should be expended on making sense of the genomic data at hand rather than working to accumulate even more redundant sequencing data faster. The recently published reports of the poor performance of well documented biomarker SNP’s for prediction of development of cancer in two separate large clinical studies seriously calls into question the whole enterprise.
The availability of a personal genomic sequence will probably generate more grief than benefit. For one thing, a very large portion of the population may find out that one or more of their parents are not who they thought they were. For another, the added stress from the psychological impact from the knowledge that one’s genome harbors perhaps hundreds of SNP’s that have been suspected to contribute to disease will probably have greater impact than the actual SNP’s. For the vast majority of people with limited knowledge of biochemistry, the actual data from their sequenced genomes will be completely meaningless except for something to brag about with their friends that they have been gene sequenced. Surely the funding to sequence genomes on mass could be better spent on more productive avenues of research that are much more likely to be translational.
@cement_head
Well, it’s no secret that 23&me are backed by Google and other VC, there is no mystery about it:
http://dealbook.blogs.nytimes.com/2007/11/26/google-backed-23andme-debuts/
Actually, $500 for a HumanHap500+ chip run sounds just around cost. Commercial providers charge about $800, and universities charge $400-600. Overhead is probably recovered from VC, as in most startups. there does not seem to be a huge underpricing set to lure people in. $500 is not cheap, after all, and would hardly cause a large number of people to spend such a sum on non-diagnostic tests, mostly for entertainment value. They are still a startup operating in a niche market, backed by VC. Yes, they are hoping to expand into the health market and make no secret about it.
Any GWAS study worth it salt should include correlation coefficients. It is still very easy to over-interpret studies, just as Brownlee has done. Can you provide a concrete example where an P values, pop% and Ptrends were not provided in an article? I would be very surprised if an article passed muster in any respectable journal without that critical information.
As for Brownlee’s writing: she presents herself as writing from an authoritative, knowledgeable position as a medical science writer, not from that of someone of the street. A science writer does not have to be a professional scientist (the best ones are not). A science writer is still expected to be knowledgeable about the field she is writing about, to the extent that she is not penning obvious errors. Brownlee is a medical science writer which has published many articles, a book and served as a visiting scholar at the NIH. There is no excuse for someone who claims to be so well-versed in medical science to have penned an article that is so full of obvious mistakes and misrepresentations.
While it’s true that genetics is only part of the equation, it is certainly a crucial part. Just how crucial depends on the disease or condition — for Mendelian diseases genetics, by definition, is the whole story (for the most part). Metabolism of drugs and other compounds is often influenced heavily by genetics. Common and complex diseases like obesity, diabetes, most cancers, etc — these, on the other hand, are unsurprisingly affected by an intricate interplay between many different factors, to which genetics may contribute less than non-genetics. But to focus on this is to do a disservice to the value that _can_ be gained from knowledge of more certain genetic influences.
As far as research goes, genetic associations may not say much concrete about risk for complex diseases but their accumulation may point to underlying disease mechanisms or avenues for treatment. Other approaches are also appropriate, and necessary, to get us there. They’re all just different tools in the toolbox.
The few studies that have examined psychological effects of genetic testing aren’t conclusive, but suggest that distress is not increased by the knowledge in the long-term and is mostly affected by underlying personality. (see http://www.ncbi.nlm.nih.gov/pubmed/19594276, http://www.ncbi.nlm.nih.gov/pubmed/18219587, http://www.ncbi.nlm.nih.gov/pubmed/18197053, http://www.ncbi.nlm.nih.gov/pubmed/19605829) Clearly, some people do not want to know, but others do, and there is likely a subset of genetic information that will not cause distress but could potentially be very useful or even life-saving.
Perhaps we can discover the genetic factors underlying willingness to learn one’s genetic data, but I’m not sure where we’d get the negative controls… 😉
Full disclosure: I currently work at 23andMe but the statements above are my own and do not represent my employer.
Also, @cement_head, I’m not sure what provoked the “PLoS and its spawn” comment. PLoS has a number of high-quality journals in addition to a high-volume journal (which still goes through peer-review for technical merit) and I hope you’re not implying that the Open Access nature of PLoS is a negative rather than a positive, or equating “high-volume” with “low-quality” and applying a blanket generalization.
Full disclosure again: I’ve reviewed papers for PLoS ONE.
@shwu
No, you mis-interpreted the PLoS comment. I meant that it is a good journal AND it publishes many SNP-disease correlation papers, THUS it appears to be a reasonable indicator of linkage (albeit with the above commentators caveats). Meaning that I don’t think that using SNPs is a bad idea. Probably more useful than gobs & gobs of non-coding sequence for which we cannot attribute a function or role. I’ve always thought that SNPs were a fairly good approach to identify physical regions of the chromosome that move about together, although it has its problems.
I actually prefer OpenAccess journals and I think they ALL should move to this model in the modern era (i.e. now). “PLoS and its spawn” as in “Nature and its spawneseseseseses”
– CH
@JeffCahn
Didn’t know that she was a science writer – thanks.
@cement_head Apologies and thanks for the clarification :). Unfortunately, not everyone speaks ironically along those lines.