Genetics Entrepreneur Anne Wojcicki Answers Your Questions


Last week we solicited your questions for Anne Wojcicki, co-founder of the “personal genetics” company 23andMe. Among your interesting questions: are 23andMe’s genetic results taken seriously by doctors? Should children have the procedure done? Will insurance companies engage in genetic profiling?

Thanks for the good questions and to Wojcicki for the compelling answers.


Have you had any situations where a person finds out they have relatively recent ancestors of a race that they were unaware of? If so, have any reacted badly? — Corey


We know that quite a few people find surprises about their ancestry through 23andMe’s service. One customer with extensive knowledge of his European paternal ancestry discovered that his maternal line traced to a Native American woman. He tracked down paper records that revealed a “mulatto woman” about seven generations back. Native American ancestry makes some sense given that his ancestry traces to the southern U.S. While he was excited about this new information, his 93-year-old mother was far less positive and remains skeptical! Quite a few African Americans have discovered that their paternal line traces to Europe. Although many African Americans may be aware that they have some European ancestry (the average is about 20 percent), some discover that close to 50 percent of their ancestry traces to Europe, and this can take some getting used to.


Let’s say one takes that test and is predisposed to a certain disease. There are always two sides: the qualitative (as in way of life) and the statistical. Has anyone done any research using both approaches simultaneously — by monitoring people who have had these tests and looking at lifestyle changes, the specific impact on life span and other particulars (like the diseases and the level of seriousness of the health problems they incur, etc.); and have they also done statistical work to see the results in a general perspective? — Science Minded


Your question is an important one. There haven’t been very many studies that look at the long-term effect of having one’s personal genetic information because widespread availability has been so recent, but these studies will be important to conduct to get a good sense of how access to this kind of information might help to engage people to be more active in their health. There have been some short-term studies on attitudes toward health and genetic information. For example, the NIH Multiplex Initiative has shown that people who voluntarily undergo genetic testing for multiple common diseases similar to what 23andMe presents are more health-minded than those who do not. And contrary to what some critics have feared, testing participants actually thought that genes played a little less of a role in disease risk than people who chose not to get tested. This result may be due to self-selection of a group health-minded infovores, or it might be because as people learn more about genetics, environment, and health, they are more open to being tested. But we will still need to study if and how people make use of their genetic information.


As projects like yours and the HapMap uncover numerous instances of genetic differences between human groups or races, what is the responsibility of the genetics community when discussing innate differences between races, particularly when a large part of academia is convinced that there are no such differences? — AC


A lot of the difficulty in talking about race has been a lack of agreement on what “race” means. In the past, the idea of pure races also included an ordering of certain races as inherently superior to others. We reject this idea absolutely. However, that doesn’t mean that there are no genetic differences between populations of different ancestral origin. A few of our features use the genome-wide data of reference populations from around the world to trace the origin of pieces of an individual’s genome. Some customers have complex patterns depending on where their ancestors originated. These reference populations aren’t “races”; they’re representative samples of peoples who have lived in a single place for a very long time and have thus accumulated different sets of genetic variants over time.


People don’t want insurance companies screening customers using genetic profiling, but do you think there’s room for insurance companies (public or private) to use services like 23andMe to assist their customers in pursuing preventative measures? People (and regulators) may not want insurance companies to deny coverage to somebody because they’re at a high risk of getting some form of cancer, but they might not mind being encouraged to get checked every year by their insurance company. Do you foresee a future where insurers might encourage clients to get genetic profiles as a cost-cutting measure? — Michael


We get a lot of questions about the potential use or misuse of genetic data by insurance companies, but many people focus on the potential for genetic discrimination. (You can see an FAQ about G.I.N.A., the federal Genetic Information Nondiscrimination Act, and how it protects against genetic discrimination on our website here.)

Your question is one we don’t get as often, and is one we have thought about a good deal. There are significant barriers to making this happen, including a general lack of trust in the motives of insurance companies and the current relatively limited use of genetic information in health care. Still, for me the short answer is yes, I think a future where insurers utilize genetic information in a positive way, to provide improved health care on a personalized basis at a lower per-capita cost, is within the realm of possibilities. Also, under a single-payer system where financial incentives reward better preventive care and insurers don’t profit by excluding riskier individuals, genetic information could be used not only to screen people at high risk for disease, but to better engage individuals to actively participate in their own health care. As the impact of genetic information in making health care decisions is better understood by scientists and physicians and as health care reform with an emphasis on preventive care and individual ownership of health is enacted, this possibility will have an increasing chance of becoming reality.


I think it is easy to see how some of the information that you provide users might negatively influence their lives. For example, they might uselessly worry about increased risks for some diseases when those risks are truly tiny and negligible. Could you please give some examples of how a typical user of your site could use the information you provide to positively influence her life? — Tim


This is an important question to address because we hear a lot of speculation about anxiety people might have when receiving their genetic information. We are not aware of any studies that indicate this concern is a reality. In fact, the REVEAL studies by Robert Green and colleagues, including the latest one published this month in the New England Journal of Medicine, have shown that people who receive genetic information showing increased susceptibility to Alzheimer’s disease did not have significant short-term psychological risks. People showing signs of anxiety tended to show these signs even before receiving their data. This result is consistent with what we have heard anecdotally from our customers. Another recent study by Colleen McBride and colleagues with the NIH Multiplex Initiative showed that people who voluntarily undertook genetic testing for multiple common diseases, similar to what 23andMe offers, were actually more motivated to change health behaviors than those who didn’t take the test. This finding counters criticisms that people will think their genetic information is overly deterministic. But generally, what we hear from customers is that their data is fascinating to them, and while there may be a few things they want to discuss with their physician, there also may not be anything they want to follow up on. Over time, our hope is that as people learn more about genetics and the role it does and does not play, each of us will gain a better understanding of our own bodies and do what we can to positively influence our health through diet, exercise, and/or intervention under the care of our physicians.


Do the results tell how much of a risk a person is at for certain diseases, or do they just show you a list of things you’re predisposed to without any clarification? Also, what is the DNA checked for? Just the basics like Huntington’s or Parkinson’s, or is it based on a health history people fill out? — Megan


23andMe currently carries information for over 100 diseases in total. For some of these, there has been enough research done that we can estimate a lifetime risk; for others we report on the latest research and how it applies to your genetics. We launched with just 13 reports and have been consistently adding reports since then. We do cover an important variant for Parkinson’s risk, and we do not cover the Huntington’s repeat.


Do you find that most people use your services for specific reasons (such as, my father died of a strange disease and I want to know if it’s something I should be worried about) or is it more about curiosity (I want to know about the amazing complexities of my own body)? — Marla


We are lucky to have a very broad range of young and old users with different interests. People use 23andMe for all kinds of reasons, from testing for a genetic predisposition to blood clots to learning which grandparent gave them their blue eyes to testing for Native American ancestry (no more than five generations back though!).


How does the service differ from Family Tree DNA? My family participated in this for about a quarter of the price. The information supplied so far is very interesting. — MB


The main difference between this service and the ancestry component of 23andMe’s service is that we use 600,000 data points from across the entire genome; most ancestry services use only a small subset: the maternal (mother’s mother’s mother, and so on) and paternal lines. This allows us to tell you more about your ancestry than any other company. We can even detect African, European, Asian, and Native American ancestry within the last five generations, no matter who you inherited it from.


As a fellow scientist, I’d like to hear your thoughts on how your work accounts for epigenetic factors. Are you looking at any gene methylation? As we learn more and more, we are finding that it’s not the sequences that are driving many disease states, but rather epigenetic controls than can be modified even in utero. — TJ


We don’t currently look at methylation or other epigenetic factors. We are tracking the interesting work currently going on in epigenetics, and it will certainly become important it in the future. However, there is still a lot to learn about the genetics of disease, traits, and drug reactions, and it’s also where most of the research has been done, so it is where we can provide the most information.


Does 23andMe have CLIA certification? Reading through the other comments, it seems a lot of people believe the reports you generate contain diagnostic information. It also seems that some people believe your scans check for single gene disorders, such as Huntington’s, that are not even SNP based. On that note, some appear to believe that you are assessing all genetic variation. What is your company doing to correct these serious misunderstandings? And what about other ethical/moral dilemmas such as testing kids’ DNA at the request of their parents, especially for adult onset disorders? Aren’t you worried that these kids will grow up to be adults who would rather not know? — Lauren


23andMe contracts the actual genotyping to a CLIA-certified laboratory; we don’t have an actual wet lab at 23andMe. It is true that while the genotyping chip we use includes hundreds of thousands of SNP’s from across the genome, the chip does not cover all possible variations, especially non-SNP’s such as large insertions or deletions. However, we do have a custom chip that presents rare variants not included on off-the-shelf SNP chips, such as the LRRK2 G2019S variant associated with Parkinson’s disease and some observers have overlooked this. We try our best to convey the content of our chip on our website; the details are all there.

Regarding genotyping of children, we have carefully considered the guidelines issued by groups such as the European Society of Human Genetics. In general, we try to give all customers, including parents, as much information as possible about potential unexpected consequences while letting them take ultimate responsibility for the welfare of their children. The guidelines have generally encouraged discussion of these issues with “competent” children who can provide some form of assent (though not legally binding consent), but they have also acknowledged that it is difficult to describe criteria for “competence.” We have chosen to give parents information and let them decide when, whether, and how to have a conversation with their children.


When do you think it will be the norm for every individual to know his or her genetic makeup at the same level currently offered by your 23andMe service? — Kay


I think in 5 to 10 years it will be standard for individuals to have access to their genetic information. As a result, individual health should improve by virtue of better preventative care as well as tailored treatments. Perhaps most importantly, the research that this large community of participants will enable will allow us to far better understand the molecular pathways of our bodies.


How does all the snickering from genetic counselors about your site make you feel? Seriously, can you guys come to my medical school and defend yourselves, because they’re really dragging your name through the mud? Doctors aren’t willing to have anything to do with the site, but there have got to be some cases where the predictive power of the results you get is really high. — Joel


While many doctors support 23andMe, individual empowerment application with respect to health information has sometimes been a protracted process in parts of the medical community. Earlier in the 20th century, cancer diagnoses were not shared with patients. Over-the-counter pregnancy tests and anonymous HIV tests both faced numerous hurdles before they got to the public. And with the arrival of the Internet, many doctors bemoaned the kind of information their patients now have access to (at times perhaps with good reason).

Nonetheless, history has come to show that these developments have improved people’s knowledge of themselves allowing them to be more involved in their personal health care. It is now hard to imagine a world where a patient would not receive her diagnosis, have easy access to pregnancy or HIV tests, or have access to the world’s vast online health information.

We believe that genetic information will follow the same course. Already, studies show that individuals with access to their APOE allele (which is strongly correlated with age of onset of Alzheimer’s) are quite capable of handling that information in a healthy way. More importantly, we believe that empowering individuals with their genetic information will lead people to be more proactive about their health and lead healthier lives.

Ultimately, we expect that all medical professionals will embrace individual access to their full genetic information and we would like to be part of that evolution.

Mike Browning

it was disconcerting to see the question about the negative opinion of most physicians about this testing show up at the bottom of the interview. It was also disconcerting to hear Ann Wojcicki comparing her product with pregnancy testing and HIV testing. Is there really any science behind this stuff? Isn't it a bit outrageous to suggest that the genetic testing she sells has predictive value remotely approaching that of HIV testing.

That this stuff is very cool gene jockey stuff is undeniable. But diagnostic for disease; in my opinion not even close. I have no brief for physicians but in this case I feel like they are on the right side of the debate.


Mike (@1), it's not *supposed* to be "diagnostic for [a] disease". It's supposed to provide information about certain kinds of risks -- as in, "Redheads are more likely to develop skin cancer" (risk statement), which is not the same as "that spot there needs to be removed" (diagnosis of disease).


Mike (#1) bring up an important point.

I suspect a lot of 23andMe's customers are not getting what they think they're getting. Let's face it, how many people know what a SNP is? How many people even realize that 23andme isn't sequencing genomes?

Questioner Megan seems to be better-informed about genetics issues than most Americans. (How many people can tell you what Parkinson's or Huntington's diseases are?) But despite being better-informed than most people, she still didn't realize that 23andme's technology cannot screen for Huntington's.

It's true that 23andme's web site does not convey any false information about what's on the chip. But It also does not convey information about what's NOT on the chip. In fact, their "Genetics 101" primer doesn't even mention that SNPs are only one type of error that can exist in a genome. No mention of insertions, deletions, or copy number variation. SNPs are all they talk about, and why not? SNPs are all they screen for.

Of course, you'd hardly expect them to broadly advertise their limitations. If genetic screening is a product, then I suppose I really shouldn't be bothered that many consumers don't know exactly what they're buying. Most people buy a car without realizing that they're getting ripped off, so why should genetic screening be any different? (Heck, 23andme costs far less than a car, so if anything, getting ripped off on a car is a costlier mistake.)


Robert A. Berman

I believe that information possibly concerning the health of an individuals body and mind should always be made available to them with enough other data and counseling to assure understanding & implications of that information.
This should be legislatively assured, with penalties for withholding such information. I believe that Anne Wojcicki is correct in noting that testing for HIV and other culturally controversial conditions has been restricted or compromised by opinion. That genetic testing is not predictive is not always true and the predictive ability of such testing (which will be more sophiticated in the future) will improve over time. One example of directly useful information for genetic testing is the recently acquired ability to determine whether a specific treatment is beneficial to a particular person -- the BRCA1 & BRCA2 gene variants are now routinely used to determine beneficial breast cancer treatment.



I'd like to know if the information is actually being used preventatively. Lots of people already know if thery are predisposed to diseases (I'm thinking specifically of obsity) and don't do anything about it, or at worst use their genetic predisposition as an excuse. Of course providing information gives individuals the choice to change their behaviour with full knowledge of the risks, but often this information is not acted upon.


Quite interesting. i was asked to participate in a genetic family study (considering cancer) and i refused. This has had negative psychological results on relationships in the family when some members were very disappointed not having my input or data.

However, I do take care of my health and have experimented with variations in diet and exercise. I have based my continuing health consciousness on my looks -weight, skin, hair, nails, etc and general feelings of youthfulness, and also reactions of other people to those features.

I prefer not to be "studied" genetically, but i guess observation of the the unstudied is a study in and of itself, too.

Ted Smith

As a customer of 23andme, I can tell you that I knew exactly what I was getting when I spit in the tube. I knew that I was not getting my genome sequenced. I knew what they could provide and what they could not. I think they actually go overboard in reminding on their site how non-deterministic the results are.
Two months ago, I went to a prenatal screening with my wife and a genetic counselor showed up to do a family history analysis with us. I mentioned to her that both of us had done 23andme, and she was appalled. "That kind of information shouldn't be made available but through a genetic counselor."
Look, I got to CHOOSE to see my SNP data. I don't need someone to hold my hand and counsel me on the implications. I'm a big boy, and I can do the research myself. I resent the idea that some "genetic counselor" should present me with his take on the data (especially when I am probably better informed than he is).
The beauty of it is, if you don't want to see your data or don't think you could handle it, you don't have to buy the service!



Genetic testing is very different from other types of testing in that it is very information rich. One aspect is knowing exactly what is being tested for and how accurate that testing is for detecting all mutations or only a specific mutation has enormous impact upon the meaningfulness of the genetic test for a patient. For example, the most common mutation in Caucasians for cystic fibrosis is the deltaF508 mutation. Hispanics and Asians (although at a low risk for cystic fibrosis) are much less likely to harbor this mutation. Therefore it is important to advise a person if they have only been tested for the deltaF508 mutation that they are STILL at risk for being a carrier for cystic fibrosis. Many individuals would go away from a test that only tests for this one specific mutation thinking "I've been tested for CF and my offspring do not have a risk for that disease" which may not be true in many circumstances.
This is why there is concern among health care providers that understand the nuances of genetic testing.



Thanks for answering my question Anne.

Yeah, I imagine a lot of doctors are welcoming the work that 23andme does. It's probably only the geneticists that snicker.

Here's why geneticists don't get along with 23andme. Geneticists only see sick people or family members of sick people. I should ask one what they would do if a totally healthy patient with no particular disease recurring among her family came into the clinic. I honestly think they'd refer her to a psychiatrist.

23andme on the other hand will test anyone.

Geneticists only offer diagnostic services. You come in with a problem and they try to tell you if it's genetic. You come in with one burning question and they give you a straight answer (more or less) 23andme isn't really there to diagnose. It's there to answer a thousand different questions that you haven't even thought of. And yeah, sometimes it is diagnostic. Take sickle cell anemia. That's caused by a SNP and as far as I know everyone with the disease has the same SNP. 23andme can tell you for certain if you and your partner will have either a 25% or a 0% chance of having a kid with sickle cell (ignoring spontaneous mutations). That's diagnostic work geneticists can't snicker at.

I agree with Melissa in the comment above me. That's exactly why geneticists don't like private genetic testing. Of course, I think that perception is totally nuts. I'm willing to bet that 23andme tells their clients straight up that every single SNP mutation can't be tested and to only put their faith in the positive predictive value of the test not the negative predicitve value. And a lot people might not agree but I think the general public won't have any trouble differentiating between positive and negative predictive values.

Besides, in a couple of years microarrays will be big enough to fit 99.999% of the SNPs for cystic fibrosis and BRCA1 and whatnot and you won't have to worry about lousy negative predicitve values.

But on the other hand, there's an episode of South Park called Cherokee Hair Tampons where all the adults in South Park get really into herbal medicine. One of the boys ( Kyle, I think) gets a messed up kidney so they take him to the herbal remedy shop for spiritual healing and the mexicans (voiced by Cheech and Chong) pretending to be Cherokees tell Kyle's mom that Kyle should really see a doctor.

23andme isn't for sick people either. Though for the sake of not being mean, I'll say that unlike holistic medicine, 23andme is good for healthy people.


Ted Smith

@Melissa - I would tend to agree with you except that 23andme specifically notifies us on the page displaying our deltaF508 result: "Please remember that the DeltaF508 mutation is only one of more than 1,000 in the CFTR gene that can cause cystic fibrosis. Its absence does not rule out the possibility that you may carry another genetic variation that causes the disease."


For those of us that have no history, 23andme is like a dream come true. Thanks to the secrecy of adoption and sperm and egg donations, there are thousands of us with no history or in many cases, lies.

I am curious, curious, curious, especially about ancestry. I feel like I fell off the family tree and I want to climb back up and find the roots and see the branches.

A new price cut and I'm good to go. Looking forward to this for a long time.

Just snapped the blimp at University and Cowper in Palo Alto the other day. My son got it at Burbank airport. Love it!


Are there any positive or negative way that entrepreneurship has affected the health care field


This comment is for #7 Ted smith. I would love to talk to you about where you got your testing done. Im very interested in getting mine done. I have been very ill for some time. My father died of cancer a few months ago, and our family history is not in good shape. Please recomment and maybe we canget in touch. Thanks. Looking for answers. Cindy.

Herbert Lewis

Did I recently hear--or imagine that I heard-a report that someone reported a genome based study that indicated that the degree of genetic difference within each of two populations is less than that between them? If there was such a report I'd like to track it down.


Herb Lewis

P.S. This is not necessarily for posting--unless there is an interesting answer.