Learning from Esther Dyson’s Genome

If you volunteer for Harvard Medical School professor George Church’s Personal Genome Project, you’d better be ready to have your full medical records along with your full gene sequence (once completed) made public. But why would anyone want that kind of exposure? Famous venture capitalist Esther Dyson explained her reasons for being one of Church’s first ten volunteers last week at Fortune’s first iMeme conference in San Francisco.

Church (who is also an Xconomist) hopes to gather enough data from the project to speed research into the links between gene variations and both common and rare human diseases, and to accelerate progress toward more individualized health care based on patients’ genetic profiles. Dyson’s reasons for participating dovetail with Church’s vision:

• She hopes that even with just 10 participants, the project will begin to generate some interesting data.

• She hopes to prove that “doing this isn’t nutty… Putting your genome up [online] isn’t the equivalent of putting up a virtual voodoo doll that people can stick pins in.”

• She wants “to ask why”—to find the genetic explanations for our appearances, behaviors, and illnesses.

Dyson, who sits on the board of West Coast personal genetics startup 23andMe, admits that it’s luck, in part, that’s allowing her to take part in the study. Her boss can’t fire her because she has a “weird gene,” and even if she loses her health insurance, she has the financial wherewithal to cope, she explained.

Ultimately, Dyson says, she hopes the choice the high-profile “first 10” have made will help push society to figure out the implications of so much genetic information becoming available—publicly or privately.

Conference video of Dyson’s (very short) talk: http://money.cnn.com/magazines/fortune/techconference/2007/.

Trending on Xconomy

By posting a comment, you agree to our terms and conditions.

11 responses to “Learning from Esther Dyson’s Genome”

  1. There are so many things to talk about with this that it is difficult to know where to begin, so I’ll pick one at random: correlation.

    I assume that Ms. Dyson knows by now that she doesn’t have cystic fibrosis or some other dread disease (thank goodness). This leaves a mishmash of other possible “hits” in her genome that may or may not have some correlation (weak or otherwise) to a different, perhaps more latent condition like Parkinson disease though hopefully not for Ms. Dyson. She’ll probably see that she is a carrier for a few horrible congenital genetic diseases. The carrier frequencies for such conditions are high in the populace but the odds that two carriers meet and make babies together is low. It is further remotely possible that Ms. Dyson’s own genome has a bona fide mutation that usually causes real trouble but for some reason, isn’t a problem for her (it happens once in a while). In short, she’ll find what every one of us would also learn were we to participate in the same project. But what does it all mean?

    We know that the “genome project” was a bit myopic. It’s a sort of reference sequence for the human genome but at the end of the day, an arbitrary one (whose genome is it?). It was an important start towards trying to understand the vast complexity of human development at the level of our genes and with any luck, one that might lead to improvements in heath care as well as a few bucks for those best able to interpret it. That said, and to quote from the movie Quadrophenia, “What’s normal then?” What I mean to say is who among us is “normal” genetically? Nobody is and everyone is.

    Correlations between our genes and health are sometimes easy to make (like in the aforementioned case of cystic fibrosis) but become significantly harder when we tackle a condition like diabetes and harder still with a subject like preference for Coke versus Pepsi. Lining up ten people to have their own genomes sequenced and compared is an interesting experiment in my mind though one that might unfortunately lead to a few sleepless nights for the participants depending upon what blemishes are revealed in their DNA. I’m looking forward to hearing how similar the ten genomes are though I expect the answer will be that they are so fantastically alike that only a geneticist would care about the differences.

    What will the information from these ten genomes tell us about drug development? I don’t think that it will tell us anything to be honest, but admittedly, it’s probably not intended to. Developing the technology to both sequence and analyze a “personal genome” quickly, accurately, and cheaply is a huge undertaking. But, is it worth it? Yes, I think so. What will it give us that we don’t have right now? I believe that the answer comes from an old description for the two types of pathologists: lumpers and splitters. Lumpers say “all of these things are alike so we’ll treat them the same” and splitters prefer to categorize based on how things differ.

    Lumpers in drug development work from the premise that starting with a particular compound, one wishes to find all relevant conditions that might benefit from its use, creating a larger market for the drug in the process. An example here would be Gleevec which was launched for the treatment of chronic myeloid leukemia but has also proven effective against certain gastrointestinal tumors. Presto, a larger market for Gleevec and a greater return for Novartis. One could also lump based on shared features in accessory or secondary pathways in disease. Lumping people into groups based on their responsiveness to certain types of therapeutics could lead to combination therapies useable in seemingly unrelated conditions. An example here might be people who metabolize certain compounds a bit more poorly than others, a difference that could be identified prospectively at the DNA level as long as the correlation had previously been made using a good data set.

    Splitters start from the standpoint of a particular disease, take colon cancer for example, and ask what differences there are between cases, say early and advanced stages that might be better served as individual conditions (and markets) due to differences in underlying etiology. Define such correlations and presto, more targets. The endpoint for all of this as far as the patient is concerned is treatment tailored to the specifics of their illness. No more “one size fits all” therapy. The endpoint for companies is perhaps a greater variety of therapeutic targets provided enough people fit into these new markets to make development financially worthwhile, even as orphan products. Now for the big question, how are these targets/new markets going to be identified? In part, via personal genomes. Here’s how…

    When a patient has a condition with an underlying genetic basis (either inherited like cystic fibrosis or acquired like colon cancer) the primary “hits” may be few. What this means is that a small set of (or just one) major “smoking gun” mutations may initiate a disease or mediate the wealth of problems associated with it. That said, if there is one thing that decades of developmental biology and cancer research has taught us it’s that nobody works alone. The types of diseases we are talking about can be complex, especially the acquired ones and the action of other genes (healthy or not) will act in concert with the “smoking guns” to shape the course of disease progression. Subtle differences between each of us that in health don’t matter much at all might influence the course of our decline when we have a disease. Being able to divine these relevant secondary players could do a great deal to shape treatment and thus, having the practical ability to scan for them early could influence treatment from day one. A similar notion is behind the quest to identify biomarkers for disease that are detectable in serum or urine. Biomarkers are related to metabolism which is of course related to genes and thus, should correlate with what we find in our own genomes.

    It’s going to take a while to see this happen. Not only does the tech have to be worked out to make having the data realistically possible, but other attendant issues like comparing many such personal genomes in health and sickness must be done. We’ll need to know what is signal and what is noise, especially in conditions like malignancy where tumor cells have mutations not present elsewhere in the body. The issue of patient insurability is huge here as well (should/does a secondary “correlation” actuarially influence risk). Who knows how long it will take for personal genomes to become useful, but one thing is for certain, someday it will be said that it all started with the genomes of ten volunteers.

  2. What a loaded issue this is! Today’s survey on the Genome Technology website indicates that most respondents think it doesn’t matter whether or not sequence data is “anonymous,” since no matter what it uniquely identifies an individual (and possibly their close relatives too) and can fairly easily be linked to other information to personally identify them. And it’s not just about sequencing – if you genotype thousands of SNPs and deposit the results in a public database, the same issues apply. To me it’s just another example of the maxim that (to paraphrase a comment I heard from John Perry Barlow of the Electronic Frontier Foundation) “there’s no privacy in cyberspace, so get used to it.” That’s why it’s so important to have legal protections in place so that people can’t be discriminated against or otherwise disadvantaged based on their genome sequence.

    Then there are the many ethical dilemmas that can arise when studying individual genomes. For instance, suppose we identify a variant conferring serious disease risk, like sudden cardiac death. Should we — or even could we — reach back through the layers of anonymization and other impediments to offer that information and appropriate counseling to the sample donor?

    It’s one thing to consider these matters dispassionately, quite another to subject yourself to them directly and publicly. For that I have the highest regard for Esther and other participants in the PGP (which will be expanding enrollment in September, btw). As a bioinformatician in a high-throughput genotyping lab, I often find myself in the midst of such discussions. I wanted to understand what it’s like at the other end of the DNA sample processing pipeline, so I volunteered for a similar NIH-sponsored sequencing project. I’m here to tell you that it’s quite a different matter when the discussion is about your genome, your family and your future. But I want to know because I really do believe that knowledge is power. Ironically, when I met with a genetic counselor for my intake interview, the one question she wasn’t prepared with an answer for was: what happens when I want my sequence? how are they going to provide it and under what conditions? I may be unusual in that I want it and know what to do with it, but they definitely have to think about it as long and hard as they thought about the many other thorny issues surrounding medical sequencing.

  3. I’m not known for brevity (sorry)…

    I must ask you Lee, was there a consent form involved in the study in which you participated and what did it say?

    Human subject research of all kinds, whether tissue samples are taken or even if only questions are asked, is almost always approved by an Institutional Review Board or IRB. The process that a public “personal genome” project would go through to be approved by an IRB is probably a fascinating tale of its own. Here’s what I’m getting to with all of this…

    I’ve been involved in several gene mapping projects (back in the old days when we did Southern blots and sometimes PCR to find disease-causing genes with the help of big families willing to donate some DNA). In every instance, the consent form indicated (because the IRB demanded) that there would be no direct flow-back of any data whatsoever to the volunteers. Sure, the families could read the results in a manuscript if one were published, but they couldn’t phone us up for their own raw data. Furthermore, they could not claim any rights to it at all. This has important implications for diagnosis as research results are not the “standard of care”. It also could mean a great deal in the realm of royalties if IP results from a discovery made using a volunteer’s DNA (and is a big topic of discussion/contention in certain circles today as to whether or not volunteers should expect a share of the wealth). Thus, I am left to ask again, what did the consent form say for your study? Did it state that you could expect your own results or DNA sequence? I suspect the answer was a firm “no” unless IRB’s have changed their philosophy.

    That said, the question remains, should you reasonably expect the primary data from your own genome to be shared with you? If you can get your own data, anyone should be able to get theirs too, right? True, not everyone is a skilled bioinformatician and would thus likely have no idea where to start should they wish to analyze their sequence. However, the internet offers a great deal of on-line DNA analysis packages for free to scientists or curious dabblers of all types. I don’t know what I’d do without them to be honest. This is where it could get very difficult as I’m sure you know.

    It reminds me of something I heard about years ago. Somebody was talking about an over-the-counter HIV test kit. Might be easy enough to make. Could even sell. But, would it be a good idea? I think that it would be a horrible idea! A false sense of security could stem from a negative result. Mistakes could happen using even a very reliable kit. Perhaps an even greater danger, what would the situation be like when someone got a positive result (real or false)? Where would the counseling be? Where would the support be? Where would the resources be? Nowhere. Maybe a pamphlet in the box.

    Somebody gets their genome sequence on a set of CDs in the mail. What will they do with it? Of course, they’ll look at it. Maybe they’ll crunch through it using web-based tools. They’ll find some of the stuff I talk about in my comment posted above if they have any idea what they are doing or want to figure it out. Maybe they’ll find a lot of repeats in their Huntington’s disease locus, which could be very bad indeed. Maybe they’ll find a lot of mutations in their glymogene locus and completely freak out (which would also be bad because there is no glymogene locus). I worry about this sort of stuff. I’ve been around the genetics clinic and elsewhere when people got results that shattered their lives. The importance of having people present who know what to say and what support and services to offer at just that moment is impossible to overstate.

    Finally, it’s true that most people would likely have no idea what to look for in their own genomes. It’s also true that your genome is “yours” so to speak though I am unaware of such a legal determination here. Maybe it’s more correct to say that your genome IS you in a way. But is a photo of you taken by a professional photographer similarly yours? It is if you agree to those terms with the photographer, otherwise the answer is no. I think that here the terms are likely that your genome is not property just as your own liver cannot be commodified; a liver actually cannot be sold under the organ donor laws in the US, though they appear to be silent where DNA is concerned. However, unless I miss my guess the interpretation and/or representation of your genome, as A-G-C-T, may indeed be subject to ownership and likely does not, what’s more CANNOT, belong to you (because the IRB would insist). That data belongs to the lab that generated it and the consent form you signed should say so clearly. That is, unless things have changed a great deal in the past few years.

  4. Jason Bobe says:

    Willy –

    I’m going to respectively disagree about tone of this one:

    “Where would the counseling be? Where would the support be? Where would the resources be? Nowhere. Maybe a pamphlet in the box.”

    How about a phone number? Video cam? House call? Those of us who would choose our homes over some institutional setting shouldn’t be discriminated against, honestly. Who says life shattering news is best received in the sterility of a hospital while wearing a paper gown? Bah! We should argue about this sometime.

    Some of your other comments lead me to believe you should totally check this article out:

    Greely HT. The Uneasy Ethical and Legal Underpinnings of Large-Scale Genomic Biobanks. Annu Rev Genomics Hum Genet. 2007 Jun 5.

    Another thing I would like to argue about sometime is the wisdom of making such an important article inaccessible to people from the comfort of their homes. OK, someone could pay $20 if they really wanted access, but is that *really* necessary? Sheesh. Fiesty tonight.

  5. Ken McClellan says:

    In some cases, the conclusions of linking genes to conditions will confirm what doctors have suspected — like one group of people having a greater probability of deveoping cancer or diabetes than another. In others, it will absolutely stand western medicine on its ear, showing us how some really bad “side effects” have been handed folks with their prescriptions because we didn’t know enough. We’re about to find out that if the first rule of medicine is to “do no harm” then our current approach of routinely handing out the 80-percent solution as the next step for a patient might not be the best way to go. The price of genetic testing needs to come down and medicine and diets need to become more and more tailored to what our DNA needs.

  6. I’m responding to Willie (sorry for the delay).

    I don’t need a consent form to know that any sequencing they do as part of the ClinSeq project is mine, in the sense that everything in my medical record is mine or at least I have the right to the information, just like any other test or imaging study or what-have-you in my medical record – it’s the law. This is very different from a pure research project, because once a potentially significant result (disease risk) is identified in part of my genome, it will be re-sequenced in a CLIA-approved lab and become part of my medical record. And in the project I’m involved with, ClinSeq, they’ve gone to great lengths to ensure that we understand what we’ve gotten into and know what resources we have available (and how to use them) to make sense of the information we’re given.

    As regards consent forms and IRB reviews and the “flow-back” of data to research subjects, I do understand the issue there. In fact as we’ve gone through consent forms line-by-line and had these discussions for GWAS studies I’ve actually heard PIs sign with great relief upon confirmation that there was no obligation to convey potentially significant results back to their subjects – it’s just very complicated and problematic and hardly anybody wants to have to deal with it. But that doesn’t change the fact that I lay awake at night struggling with what to me is a very real ethical dilemma.

    Now, what people would or could do with that information is, as you note, a very important issue and becoming more so every day as more unregulated genetic services are offered to anyone who can afford a few bucks and figure out how to do a cheek swab. It will be very interesting to see how it’s handled by mainstream vendors like Illumina as they pair up with the new generation of “personal genomics” start-ups like 23andMe!

  7. fam says:

    This is very different from a pure research project