Skip to content
Who's in the Video
Paul Hoffman is an award-winning journalist and biographer. But there are 168 hours in every week, and so he is also a budding restaurateur (at Rucola in Brooklyn) and evangelist[…]

Journalist Paul Hoffman moderates a 50-minute panel discussion about in-home DNA tests and the connection between genes and disease.


Harry Ostrer - Director, Human Genetics Program, New York University School of Medicine.

Esther Dyson - Entrepreneur, Philanthropist, Board Member of 23andMe.

Boonsri Dickinson - Assistant Editor, “Discover Magazine.”

Tara Matise - Computational Geneticist, Rutgers University.

Panel moderator: Paul Hoffman -Editorial Chairman, Big


Paul Hoffman: Let me introduce all the panelists starting with Esther Dyson, she is an entrepreneur, she invested or a board member of perhaps 20 companies, high tech companies. She is most importantly, she’s my Facebook friend, and if you saw her site recently, she was wondering about whether she’s traveled too much. Because the safari web browser shows that the United Airlines pay-for-your-itinerary page as one of her most common web pages.

But she came from California earlier this week, she’s going to California tomorrow, she was in Moscow, you heard the President [Barack Obama] speech I understand.

But the reason we have her here is for a bunch of reasons. She’s one of the few people in the world to have her genome sequenced as part of the project out of Massachusetts. By our estimation, only fifty people in the world have had their genome sequenced at this point. And she’s also on the board of directors of 23andMe which is one of the personal...


Esther Dyson: ...sleazy direct-to-consumer marketing companies. [Laughs.]


Paul Hoffman: I see we’re going to have a lively discussion. [Laughs.]


Harry Ostrer: ... I did not use that... I did not use that adjective... [Laughs.]


Paul Hoffman: So we’ll get to her in a moment to talk about that.

Next to her is Boonsri Dickinson, she is an editor of “Discover” magazine. I came across her work because she wrote a very interesting article where she had genetic testing by three of these companies [Navigenics, 23andMe, deCODE genetics] and wrote about the results. We will talk about that in a minute.

And then Tara Matise is the last of our panelists. She’s an associate professor at Rutgers University, she was part of the human genome project, she’s done work at various points in her career on the genetics of schizophrenia, she’s working now, on long term way, on the genetics of fertility.

Esther let me start with you for a moment, what was it like to have your whole genome sequenced?


Esther Dyson: Honestly, not that exciting. It was frustrating. First of all, it took a long time for them to get around to doing it. They took some blood, then I waited a long, long time. And then they told me, "Gee, you know, there's not anything really exciting." And as Dr. [Harry] Ostrer said, I could have told them that without $400, which is now the retail price at 23andMe, or not $1,000, or not even several hundred thousand.

The fact is, those of us who are interested in doing this for ourselves, we're not interested, many of us, in finding out about ourselves, but in offering data back into the pool so that in the future other people can find out about themselves.

Today I'm 58 and I'm not dead yet. So it's pretty clear there's nothing drastically wrong with me. I may get Alzheimer's, I probably won't get Parkinson's, it's not in the family, neither is Alzheimer's. So as you [Dr Ostrer] said, knowing the family history is pretty indicative.

But in the future, the data that 23andMe is collecting, and that in the mean time it shows to its customers, because I believe that's their right, is going to be very, very useful. Over time we'll collect more data, we'll collect a lot of phenotype as well as genotype data. And our pitch to our consumers is, we try very carefully not to present it as a medical service and it's not with a wink. It's because it really isn't. It's because it's interesting, it's fascinating, you can call it narcissism. Some people are interested in their ancestry, some people are adopted or for whatever reason don't know who their parents are. And for them, I think it's a lot more fascinating than for most of us.

But the real goal here is that the scenario you [Dr Ostrer] described at the end, the 61-year-old man, to help make that happen because it's going to require hundreds of thousands of genomes and of phenotypes to give us that kind of information. And we think getting consumers excited about is the best way to do that.


Paul Hoffman: Have you found stuff about your own ancestry? I understand you’ve had relatives do this.


Esther Dyson: Well, I've had 30 relatives do it. It's fascinating. But medically useful? No. There's a chance it could be but honestly, knowing that you're--I think the advantage of knowing you have a quote "24 percent risk" of something, which I agree with you [Dr Ostrer] is a spurious number, it's not necessarily that the advice is going to be any better, but looking at that number makes you take it more seriously. It's psychological. But every doctor knows how important psychology is.

Compliance is probably one of medicine’s worst problems. It's getting people to do what they know they should do already. And if looking at your genome and understanding your risk of type 1 diabetes is something like 70 to 80 percent heritable, and your risk of type 2 diabetes is only 25 percent heritable, which means the rest of it, the rest of your risk of type 2 diabetes is what you do. That kind of quantification, not the precision, but the quantification, I believe can help to modify people's behavior.


Paul Hoffman: Great. We’ll have a more free wheeling discussion about this second.

I’m just going to go through each of the new people that have joined us and get a little more about their story and why they’re here.

So, Boonsri, tell me about the three tests that you took. I understand you came back and you started worrying about Crohn’s disease and mascular degeneration.


Boonsri Dickinson: Two of the companies, 23andMe and deCODEme, gave me some information that I could check, like, when I drink wine, my face gets flushed, I have dry earwax, I’ve brown eyes, I’m not a sprinter. And although it said that I’m likely tolerant to lactose, if I had a sip of milk, you guys would all want to hold your nose.

When I read my report on common diseases, I realized the information is useless. There’s no way to tell how accurate the so-called risks are until you actually have a disease.


Esther Dyson: Even then you don’t know what the risk was.


Boonsri Dickinson: Right. Crohn’s disease and macular degeneration are two diseases that I pay attention to, because I have poor eyesight. So loss of vision is something that meant something to me. And then Crohn’s disease is another one. I have a sensitive stomach. So I’ve learned what foods to avoid.

But the limitations, for me, were far greater. I’m half Asian and half European. So when I looked at the profiles, I got conflicting results just because they ask if you’re European or if you’re Asian. They haven’t worked out a way. Although one of the board members said that he was going to work on that.


Esther Dyson: Is that Patrick Chung?


Boonsri Dickinson: George.


Esther Dyson: Oh he’s, sorry, scientific board.


Boonsri Dickinson: And so they haven’t found a way to calculate risk for people of mixed ethnicities. And since I wrote this story [“How Much Can You Learn From a Home DNA Test?”; from the September 2008 issue of “Discover” magazine] a year ago, the companies continue to update their Web sites with more diseases, more risks. But just because they’re adding more data doesn’t mean that they know more about the role that genes play in the onset of disease.

And only, Navigenics, the third company that I tested with, they offered in-house counseling. And even then their genetic counselor wouldn’t really tell me what my genetic profile meant. She just said I have to interpret it.


Paul Hoffman: So I’m curious, you’re very young. You were 25 when you did this [had genetic testing by Navigenics, 23andMe, deCODE genetics]?


Boonsri Dickinson: Uh-huh.


Paul Hoffman: Have you altered your lifestyle in any way because of the results of these tests?


Boonsri Dickinson: So I’ve changed a few things. I wear sunglasses now and I eat spinach. And my friends to try to get me to drink less, and now I do. And now they’re always begging me to drink more.

I think the whole experience has made me think more about my health. When I was 20, I got really sick and I just ran from the possibility that I was actually sick. So now, when I get the cold I’ll run to the doctor. I certainly think more long-term about my health.

When I wrote this story [“How Much Can You Learn From a Home DNA Test?”; from the September 2008 issue of “Discover” magazine], my editor wanted me to write it in the first person because I’m young. And when I handed her my story, her eyesight began to fail and she can hardly read the words that I wrote and over the next few months, I watched her get really sick. So that experience has made me think more about my own mortality and the diseases that are presented in these tests, suddenly became more real to me. And while it scares me, I can’t live my life in fear.

So I wouldn’t really recommend that you guys go spend your money on these tests. They’re not ready for primetime. But if you want a way to lead a healthier lifestyle, then maybe this is one way to do it. But it does have a psychological effect. So you just need to be aware of how you’re going to react to the information. How you deal with it.


Paul Hoffman: So Tara, when I looked at you profile in the Rutgers website [], it says you’re a computational geneticist and you pointedly said that you don’t work in the wet lab. So I take it that means you don’t deal with anything gooey. Tell me what it means to be a computational geneticist.


Tara Matise: No squishy, no gooey, just hard keys. That’s exactly what it means. I do my research entirely on computers. I’m involved in a number of studies that aim to identify genes for specific diseases. And I’m the person who is doing the data analysis mainly, or I might be writing a computer program to help perform that analysis. I might be using programs that help to automate analysis and keep things efficient and streamline. That’s pretty much what it means.


Paul Hoffman: So tell me about you work on the genetics of fertility. What are you trying to find out?


Tara Matise: I am running a study to identify genes that play a role in determining when women will become infertile. This is the genetics of female reproductive aging, which is a process that all women do go through, obviously. And on average by the time a woman is 51, she has reached menopause and approximately 10 years before that time, is a state of less fertility, and 10 years even before that time is the beginning of sub-fertile period. So about 20 years before you’re going to reach menopause, you are already at reduced fertility, and there is obviously some range of age at which women will reach menopause. And some women get there faster than others.

So I’m interested in understanding how genes play a role in determining how quickly a woman will go down that path. Ultimately the goal would be to identify genetic biomarkers that could be used to give a woman her profile and say, you have the profile of a woman who’s likely to reach infertility at a later than average age, at an average age, or at a younger than average age. And that might be useful information to some people for family planning.


Paul Hoffman: Have you had personal genetic testing?


Tara Matise: Coincidentally I’ve participated in a study that was done through the Scripps [Research] Institute in conjunction with Navigenics, one of the companies that has been mentioned here.

I really did it more out of curiosity. Of course I was wondering what might turn up. The way that project works is it’s actually a scientific study, funded I believe by the NIH [National Institutes of Health] to look at the impact of the results of this consumer testing on individuals lifestyles. So they have you fill out detailed questionnaires before and after you do the testing, and they plan to follow up over time asking you things like: What is your diet like? What is you lifestyle like? Do you exercise? Are you a nervous person? How is your sleep habits? And then they’re going to come back and see if those things change after you receive the results of this testing.


Paul Hoffman: So I understand it showed that you were at risk for obesity?


Tara Matise: Obesity was one that came up. I don’t have a strong family history of obesity, I’m not yet showing signs of obesity although it can strike at any time and it could come. But it’s obviously something that everybody wants to watch out for.

I don’t think I learned anything from the test. I didn’t change my lifestyle any after having done the testing. But as a scientist I did find it interesting to get those results.


Paul Hoffman: Dr. Ostrer, I have a question. I think a lot of people in general are confused about what it means to have genetic susceptibility to a disease, so maybe we could talk about a range of things. Huntington’s disease, as I understand it, if you have the genetic marker for that, with almost certainty you’re going to get the disease?


Harry Ostrer: That’s true. So there are certain conditions that are so called monogenic conditions, where you inherit one or two copies of a high risk gene, and that puts you at high risk for developing that condition. So for Huntington disease, the lifetime risk of developing the disease, if you inherit a mutation, is on the area of about 90%.


Paul Hoffman: About 90%?


Harry Ostrer: Yeah.


Paul Hoffman: What other diseases are in that category?


Harry Ostrer: So Hypercholesterolemia, so called Familial Hypercholesterolemia, falls into that sort of category. And of course with these conditions, most likely one of your parents will have the condition as well, and it will be traceable through your pedigree, so that’s why I suggested before that, knowing your family history would be useful.

For about 20% of the population that’s at high risk for breast cancer based on family history, they will have a mutation in one of the two high risk genes, BRCA1 or BRCA2, and if you inherit a high risk mutation then you’re likelihood of developing diseases is in the order of 80%.


Paul Hoffman: So 80% if you have that form or...


Harry Ostrer: 80%. That’s correct. If you inherit this mutation, yes.


Paul Hoffman: What advice do you give to a woman, if you find out that she has that mutation?


Harry Ostrer: So this highlights why having individualized genetic counseling is useful. Because it’s rather tricky, and it depends on how old the woman is, whether she’s had children, whether she wants to have children, whether she’s used fertility drugs, whether she has cancer at the time of her diagnosis.

For instance, for a woman who is in her late 40’s who’s had children, who is newly diagnosed with breast cancer and is found to have a BRCA1 mutation, she in fact may decide to have both of her breasts removed to reduce the risk for developing a second breast cancer. And not only may she decide to have her breast removed, but she may also decide to have her ovaries removed, because she’s at 40% risk for developing ovarian cancer, and currently there’s no effective screening tests for identifying early ovarian cancer.


Paul Hoffman: How about prostate cancer?


Harry Ostrer: We don’t have a genetic test for prostate cancer. And prostate cancer, by and large, is a more indolent disease in men than breast cancer is in women. So, I regret to say it, but you or I or Paul may have prostate cancer, may not know it, and at our ages we may still go on to a completely normal longevity.

That said, about a sixth of the male population that has prostate cancer has the aggressive form of prostate cancer, and currently we don’t have any genetic test or test of the tumors themselves that will say, you have the mild form of prostate cancer, I have the aggressive form of prostate cancer.


Paul Hoffman: I understand what some of these direct-to-consumer testing companies like 23andMe, that there are genealogical groups that encourage you to post your genetic results so that you can compare them with other people. Is there any downside, in terms of privacy? And is there ways that this information can be misused?


Esther Dyson: Well I would argue a couple of things. First of all, it’s people’s own data, and they should have the right to do with it what they want.

This is not a perfect world. Sometimes people make mistakes, they change their minds. We used to have a world in which you could not get an HIV test without a counselor there.

We [at 23andMe] just believe we tried to give people adequate warning, we don’t want unhappy costumers, honestly. I just think it’s peoples own data.


Paul Hoffman: Right. I understand its people’s own right. What is the legal situation, Dr. Orster, in terms of how this information--do you have to disclose to your insurance company? If I get a genetic test tomorrow...


Esther Dyson: Your insurance company doesn’t care, as far as they’re concerned. First of all, it’s toxic, because the answer to this question is, right now, it’s illegal/unclear. They do know your family history, they can look at you, they want to know your body mass index, they want to know how old you are. I’ve tried this. I asked an insurance company when I was looking for insurance, no thank you.


Paul Hoffman: What’s the legal situation now, can an employer refuse to hire you if they find out...?


Harry Ostrer: No. One of the great accomplishments of the [George W.] Bush administration was the passage of the Genetic Information Non Discrimination Act which limits the access of others to your genetic information, for health and employment issues. And many states such as New York have had a quite protective legislation but, GINA as it’s called, was very successful at closing many of the loopholes.

Now someone who is going to have a genetic testing at 23andMe is probably not going to have it paid by their insurance. They probably wouldn’t seek reimbursement from their insurance, and the insurer would probably not pay for the test because it’s not clinically valid.

So therefore the issue of the insurer having access to the information probably isn’t such a significant one.

That said, 23andMe as I understand it, does testing for medical conditions as well. And so whereas an individual may be making a choice for herself or himself to post the information, in fact we are not individuals in this world, we come from family units. So in the process of disclosing personal genetic information, we’re disclosing it for our family members as well who may not be consenting to the process. And in fact what we find in the realm of medical genetics is frequently family members will disagree among themselves about who should have which genetic information and how it should be used.


Esther Dyson: Yes certainly people have a responsibility to their family just like you have a responsibility to your wife not to... talk about lots of things you probably know about that other people don’t. The other thing is, in the long run, I think insurance companies will and should want genetic data for treating people. This is what we’re talking about. In the long run it can indeed be very medically useful. But what 23andMe is doing right now is not a test or a diagnosis. It does give you some limited information about likelihood of something but that’s very, very far from the diagnosis.


Paul Hoffman: How close are we to linking genes to behavior? Is there a gay gene? Is there any evidence for that?


Harry Ostrer: There was a much publicized gay gene. A literary career was built on the gay gene, but whether it is a bona fide scientific finding seems quite dubious.

One of the hallmarks of scientific discovery is replication. And so in fact, when we observe that genes have some effect on some trait, we want to find that it’s not only present in, you know, you [Paul] and Esther, but would like to find that in fact we could test some other population and find that it’s present in them as well, with a reasonable degree of certainty.


Paul Hoffman: Are there behaviors that you can point to that for sure have a very strong heritable component to them?


Harry Ostrer: Most definitely. And those are based much more, on doing twin and family studies that specifically look at heritability.


Paul Hoffman: So what would those be? What would be some behaviors that you think are primarily genetic?


Tara Matise: I can name one. Because I asked my colleagues at Rutgers who study behavioral genetics and they tell me that novelty seeking is a behavior for which there’s a pretty well established association with a specific gene. And that was the only one they could come up with. And they’ve been looking for lots of others but they’re all very complicated scenarios contributed to by the input of many genes plus environment so...


Harry Ostrer: Now are you including mental illnesses in your characterization of behavior?


Paul Hoffman: Sure.


Tara Matise: Okay. We can talk about some others. Autism--there are a number of genes...


Paul Hoffman: Austism, there’s a genetic component to?


Tara Matise: some small percentage of cases. About again checking my crib notes from my colleagues there at Rutgers...


Harry Ostrer: At least 5 percent. Right.


Tara Matise: About 10 to 15 percent of autism cases can be ascribed to some genetic defect which is understood. But the rest of the cases remain--


Harry Ostrer: Unexplained. Which doesn’t mean that genetic factors won’t be identified in the future. But that’s the current state of knowledge.


Esther Dyson: Hence we may well therefore discover that syndromes appear to be the same thing, like autism, or actually two different diseases that...


Harry Ostrer: Absolutely.


Tara Matise: Autism probably is an example of that.


Paul Hoffman: Are there genetic associations now to traits like athleticism or physical attraction? There’s preliminary data on this?


Harry Ostrer: So, the physical attraction is the “sweaty t-shirt” study from Norway or Denmark. I read that study and that seemed rather dubious to me, even though it was published in a rather high quality journal.

23andMe, I guess does testing and predicts whether people have a predominance of fast twitch fibers that may make them more athletic.


Esther Dyson: What we do is we detect a presence of a SNP [pronounced “snip”] that seems to be correlated to either...



Paul Hoffman: Tell us what a SNP is.



Esther Dyson: A snip is a single nucleotide polymorphism [SNP]. It’s a genetic marker, basically. So these have an association, there are studies with a thousand or several thousand people and so forth, and we show the reliability of this study. Do you have an above normal chance of having fast twitch or slow twitch muscles, or something like that. And there’s so many other factors there, but yes that’s one of the things we’re looking at.

We’re also trying to discover if there’s any, if there, we would like to find out if there’s any markers associated with, whether you prefer Tylenol, what is that, Ibuprofen, Paracetamol, or plain old Aspirin. These are kinds of things that are not terribly important, but are really interesting to people.

We would like them to understand the statistics of genetics as well as they understand the statistics of baseball, and I think this is a good way to start.


Harry Ostrer: So what 23andMe is doing then, Esther, is conducting observational studies. And you’re conducting studies that are similar for instance to the Nurses' Health Study or the Physicians' Health Study, where you’re asking people to report what their behaviors or preferences or diseases are--but how do you know if the people are giving you accurate information?


Esther Dyson: We don’t. And we don’t claim that they are.

We’re also working in conjunction with, quote “real medical institutions” to do much more serious clinically validated studies. There’s a real difference. I completely agree with you.


Paul Hoffman: We’re going to open up now the discussion to the members of the audience. Any questions?


Harry Ostrer: So you want to talk about James Watson’s genome while we’re waiting?


Paul Hoffman: Yeah. James Watson, as you all know, is the man who co-discovered the double-helix structure of DNA in 1953. He recently had his genome sequenced and...


Harry Ostrer: It was published in “Nature,” top tier journal.


Paul Hoffman: Right. He found out he was at risk for early dwarfism, that he was at risk for blindness and deafness and that he was at risk for a certain muscle cramping disease. And when I saw him a week and a half ago at the World Science Festival, he seemed to be quite tall and walking and seeing.


Audience member: So since we’ve sequenced the genome, are you disappointed with the advancements we’ve made with regard to medical treatments, etc?


Harry Ostrer: No. I personally am not disappointed at all because it’s a wonderful tool.

And in fact, right now I’ve just started working with a pediatric endocrinologist in Ecuador who has discovered what’s probably a new genetic syndrome of dwarfism and diabetes within a group of interrelated people who live near Quito. And I think that the time to identify what the genetic defect in these people now could be accomplished within a matter of months, whereas in the past it might have taken years to identify what the specific genetic defect is.

So we have these really wonderful tools that make things possible that were unthinkable 10 years ago before we first sequenced genomes.

And also, you know, where we differ is in the utility of genetic information. Now I don’t think that those of us representing the extremes of this panel disagree that there isn’t going to be a place for genomic analysis and personalized medicine in the future. The issue is when should it start being offered to people. And even in my scenario, it’s quite optimistic. I think that within five years we will some have multi-genic test that are based on common variants and rare variants that some of us have, that put us in increased risk of those diseases. And we’ll be able to make statements with a fairly high degree of certainty, that indeed you’re at risk for prostate cancer, but fortunately for you it turns out to be the benign type.


Esther Dyson: What we’re trying to do is not simply say, oh you have such and such a risk of such and such a disease, but to understand how the genes actually work. It’s not simply, do you or do you not have this gene, but how does your DNA actually work? Whether are the pathways, the sequences, different proteins interacting. There’s a lot more to it than simply, you got this gene, you have X-percent chance. You have lots of different genes interacting and giving you different susceptibilities. Understanding how a disease unfolds, the interaction of your body matter, your cells, your different kinds of cells, the different peculiarities they have with the chemicals you eat, the chemicals you encounter, other effects of your behavior.

So it’s a lot more complicated than just math. There’s a huge amount of molecular biology, cell-cell interaction and stuff like that. So the job isn’t done when we have all the data. Then we need to figure out how these things interact.


Paul Hoffman: Any other questions?


Audience member: A few years ago I read a comparative study between the human genome and, I believe it was a chimpanzee. They mentioned that there was only four hundred genes differ. I don’t know if any of the panel has any thought.


Tara Matise: I think to a lot of scientists that was not particularly surprising-how similar we are--and sort of reinforced what we can see about the similarities. But I’m sure there were some very relevant findings specific to given diseases and I’m actually not familiar with...


Harry Ostrer: Right. The finding that’s most commonly sited is for a specific gene that affects language production that’s actually present in humans but not present in chimpanzees. Much press has been made about that particular gene. Undoubtedly there are others as well.

Remarkably, I don’t know many geneticists who are much taken with the comparative genomics of humans and higher primates. The thinking was achieved by sequencing the genomes of higher primates, it would be useful information for testing hypothesis of the type that you’re asking about, but remarkably it doesn’t seem to have particularly caught on.


Audience member: So I’m quite interested and curious about that novelty seeking gene. What do they know about it, what role does culture or environment play in expressing the gene, if any, and how strong predictor is it of that behavior?


Paul Hoffman: Well those of the panelists, everybody who has make up on, those who go off for a really great night on the town obviously have the gene and those who go home don’t. But seriously ...


Tara Matise: I actually don’t know most of the details. But the gene is Dopamine D4, which has also been associated with addictions and a few other behavioral traits actually, but not as strongly as with novelty seeking. That’s all I know about it.


Harry Ostrer: And not surprisingly those people don’t commonly come to genetics clinics which are not particularly thrilling places. I don’t know that much about it either I’m sorry.


Paul Hoffman: We have time for one more question.


Audience: A question for Dr. Ostrer, 23andMe is now offering high penetrance breast cancer mutations for the Ashkenazi population, and several of the companies are returning results for ApoE4. Or if you have two copies you are at 15 fold increase risk for Alzheimer’s. So it seems to me that those sorts of things satisfy all three of your criteria [analytical validity, clinical validity and clinical utility], so what’s wrong with that?

The zero dollar genome. And you have been quoted recently as someone who anticipated the inevitable driving down of prices.


Harry Ostrer: First of all, I’m amazed that they can do it because both of those genes were patent-protected. And we initiated a suit against Myriad Genetics for testing BRCA1 and 2, and we haven’t won our case yet, but stay tuned.

The problem with that model is that there’s no pretest genetic counseling so that people understand what the test is and what possible results might mean. Personally I think that any properly prudential laboratory should be able to offer genetic tests and be paid for it provided that people get pretest and post-test genetic counseling.


Paul Hoffman: Thank you very much for all attending our first panel. The human genome has been sequenced. I’m still waiting for the coffee genome to be sequenced. People are working on this so we can get a better cup of coffee. And we’ll have a 15-minute break now before we have our second panel on swine flu and the next pandemics.


Harry Ostrer: So is this what we call pearls before swine?