A case study in personal genomics

I have no strong family history of any disease, despite having 7 blood aunts and uncles and countless cousins. So when I sent my spit off to 23andMe at the start of the Genomes Unzipped project, I was expecting something very similar to Caroline’s experience: a 5% increase in risk here, a 2% decrease in risk there, nothing that would really tell my anything about my health.

However, this was not my experience. Along with a pretty interesting Y haplogroup, I also had three unexpected and potentially worrying health results. I am a cystic fibrosis carrier, a hemochromatosis compound heterozygote, and have a strongly elevated risk of age-related macular degeneration. This cocktail of genetic disease certainly was not what I came to the test expecting!

After some thinking, I decided to take my test results to my GP, and see if there was any advice or testing he would recommend. In the end, my GP referred me to a clinical geneticist, which started a cascade of appointments which in turn led to a number of important changes in how I treat my own health.

What was most interesting is how the whole experience got me thinking about my health as something I am in charge of. I have since made a number of important life-style changes, some of them directly related to my genotyping results, some more generally to improve my overall health.

The point of this post is just to go through some of the experiences, what I have learned about specific conditions, and what changes I have made to my life since. In some sense, I feel like my experience is a case-study in what good outcomes can come from personal genomics, both for specific conditions, and more generally for how genetic data can change your own approach to your health.

Meeting the doctors

My GP seemed interested in the results, and in why I had ordered them, and what I was hoping to learn. However, he said that he didn’t feel qualified to talk about the results in detail, and that he wanted to forward them to the Clinical Genetics department at the hospital. I was careful to stress that I wasn’t pushing for a referral, and that I didn’t want to waste the NHS’s time and money if he didn’t think it was worth it, but he seemed genuinely keen for me to see an expert.

The appointment with the clinical geneticist was far more detailed than I was expecting. It involved taking full family history, which didn’t show anything to be worried about. She went through all the conditions, explaining what could and couldn’t be done, and what I should look out for in the future. She was most concerned about the cystic fibrosis mutation, and spent a while talking about options for preconception testing, prenatal testing and pre-implantation screening, eventually telling me to make another appointment if and when I decided to have children.

I then had a follow-up appointment with my GP, who ordered a few more tests related to the conditions. We also talked about my overall health, did a weighing, and ordered a battery of other tests related to my weight. I also requested to be entered into the surgery’s weight loss program (more on this later).

Neither my GP nor the clinical geneticist had ever had a patient with a direct-to-consumer genetic test before (though my GP seemed to have encountered various scams). Both were glad that I had come to them, and neither considered my appointment a waste of time. My GP seemed very positive about the whole experience, and said that he thought it was a potentially value experience for a patient (on the other hand, my GP also manages to seem pretty positive about everything, up to warts and rashes, so take that as you will). Conversely, while the clinical geneticist thought there was value in this testing for people like me (i.e. people who already know a lot about genetics), she was not pleased with the idea of lay-people having direct access to the information.

Hereditary hemochromatosis, or the hidden recessive disease

My parents both carry different mutations in gene HFE, which can cause hereditary hemochromatosis, a recessive Mendelian iron-overload condition which can cause organ damage (a disease of which Kate has experience). I inherited both copies, and so am a compound heterozygote for the condition. As recessive diseases only manifest when a mutation is inherited from both parents, these often show very little family history, and can come as a surprise.

Around 10% of people who developed HFE-related iron overload have the genotype that I have, which sounds serious. That said, a bit of digging shows that in fact around 2% of Europeans are compound heterozygotes for haemochromosis. While the risk of developing iron overload is strongly elevated in these people, a bit of quick maths shows that I still only have around a 1% chance of developing serious symptoms.

As it happens, hemochromatosis is very treatable and preventable. I have seriously reduced my red meat consumption (this’ll also help with AMD, see below), and I am committed to regular blood donation with the NHS Blood Service, which stops your iron getting too high (both good things to do anyway). I am also having 5-yearly blood iron measurements. The doctors say that this combination of prevention and screening seriously reduces the risk of developing harmful symptoms to very low even for people with full-blown iron overload, so in my case it should drop my already-low risk to basically zero.

Cystic-fibrosis, or the unexpected carrier mutation

As I have already talked about elsewhere, I carry one copy of a mutation that causes cystic fibrosis. You need two copies of this mutation to show symptoms, so if I had children with another carrier, each child would each have a 1 in 4 chance of developing the disease. Most Mendelian mutations are pretty rare, but actually around 1 in 25 Europeans carry the cystic fibrosis variant, so the risk is not too small to be ignored.

The answer to this is relatively simple; the clinical geneticist said to bring along my partner for testing if we decided to have children. We could also do a more extensive carrier screening for lots of other known mutations, e.g. through a service like Counsyl. This would tell us if there were any recessive diseases (including cystic fibrosis) that we both carried mutations for. If there were, we would have the information to make informed choices.

Age-related Macular Degeneration, or the vagaries of inheritance

One major risk that jumped out of my genome was a 29% chance of developing Age-related Macular Degeneration (AMD), a progressive loss of vision in the elderly. This is compared to around 7% of the population as a whole, so this is both a high relative and a high absolute risk.

I had no family history of this disease. If you look at my parents’ genotypes, you find that each has a mixture of protective and risk variants, giving them close to average risk. By chance, I inherited almost all of the risk variants, and almost none of the protective variants, giving me a highly elevated risk. This is a reminder that, while family history can be predictive, the vagaries of inheritance can always throw up sporadic genetic risk that could not be anticipated.

There isn’t much that can be done about age-related macular degeneration. Smoking is a risk factor, and I don’t smoke, so that lowers my risk. There seems to be a weak effect of red-meat consumption on AMD, and I have reduced the amount of red meat I eat to a few portions a week (which may knock a few percent off my risk). Obesity is a risk factor as well (see the next section). There are also a few more well-replicated genetic variants that aren’t reported on by 23andMe, and these also bring down my risk a little to closer to 23%.

So, while they are certainly worth while, no amount of lifestyle change is likely to have a massive effect on my AMD risk. New treatments are becoming available (the NHS lists a number of them), and by the time I would develop the disease there will almost certainly be more. AMD is a degenerative disease, and most treatments only slow down the disease, so early diagnosis can be important. The clinical geneticist told me that most opticians would not be trained to notice the signs, and that I should see an opthamologist as soon as I started seeing symptoms.

Weight loss, or the general benefits of personal genomics

At the end of last year I weighed 120kg, with a BMI of nearly 34, rapidly approaching the “severe obesity” category. I had attempted dieting in the past, but this mostly correlated with episodes of depression and self-loathing, and in the long run only made things worse. For a long time this felt like a fact of my life, something that I was powerless to change; I had more or less resigned myself to developing diabetes and heart disease. From talking to other people in a similar situation, this feeling of resigned fatalism actually seems like a pretty common approach to people’s own health.

However, as I started to engage with my health in other ways, my approach to my weight changed too. My weight, my blood pressure and my lipid levels, like my genotypes, were interesting facts about my life that I could act on, rather than depressing statistics forecasting my early demise. In fact, when I learned that I was showing attenuated liver function, probably as a result of fatty liver disease, I treated this as actionable medical fact, not as a depressing judgement.

My GP enrolled me in a weight loss program at that surgery. I started taking a food diary, altering my eating and purchasing habits, and doing more sport. To cut a long story short, I have since lost around 11kg, bringing my BMI down to 30, and my blood pressure looks healthier than ever.

From case study to population health

This post has taken me a long time to write (thanks for getting to the end!), and it is now well over a year on from when I initially got my results back. Part of the reason for this delay was that I wanted to make sure that my lifestyle changes were going to stick, once the initial buzz had worn off. Touch-wood, I am currently pretty confident that I have made lasting positive changes to my behaviour.

My own experience is only a case study, and is hardly evidence that direct-to-consumer genetics can reliably induce beneficial outcomes. The evidence from current research, such as Theresa Marteau’s Cochran review of out-patient genetic testing, and the Scripps study of behavioural change amongst personal genomics customers, seems to show very little change in health-related behaviour (though Marteau’s review did show evidence of that genetic data return can lead to improved dietary data). Currently, there isn’t much good evidence on how personal genomics can drive health outcomes, and virtually none on the differential benefits of testing for different conditions and different ways of communicating the results.

I suppose my own experience, and the similar experiences of others, illustrates the potential positive impacts that personal genomics can bring. Personal genomics, and especially things like direct-to-consumer genomics where you are fundamentally in control of your own data, can help people engage in health, both specifically for conditions that they are at risk for, and more generally for their overall health. In my own (possibly extreme) case, it transformed health from something that passively happened to me into something that was I could learn about and change, a positive (and interesting) aspect of my life. I don’t think this is a necessary, or even particularly common, outcome of personal genomic testing. However, I hope that new discoveries, technologies and methods of communication can allow more and more people to become engaged in their own health.

  • Digg
  • StumbleUpon
  • del.icio.us
  • Facebook
  • Twitter
  • Google Bookmarks
  • FriendFeed
  • Reddit

22 Responses to “A case study in personal genomics”


  • Moreno Colaiacovo

    Thanks for sharing your experience with us! Me too I’ve discovered something that was unexpected before joining 23andMe. I think it’s important to highlight this fact to potential new 23andMe customers, because they must have clear in their mind that they could really discover something weird. However, your story shows also that we can do something, even when we receive an unexpected bad news.

  • The claim that the Bloss et al. NEJM paper (the Scripps study referred to) found no effect on behavior change is not correct. Participants with higher risk estimates for obesity had higher fat intake, and decreased exercise, at follow up, so of anything it made their behavior worse possibly because those participants were succumbing to genetic fatalism/determinism. It should be (but isn’t) stunning how misrepresented that paper has become. Are people willfully misrepresenting it or just not reading beyond the abstract? Then again, it’s only a 3 month follow up so might not actually tell us much at all of any value.

  • @Jonathan

    In my defense, that finding it not mentioned at all in the paper. There is a different between “not reading past the abstract” and “not reading Supplementary Table 6”. The authors actually say

    There were no significant associations between composite measures of risk and follow-up scores on anxiety level, dietary fat intake, and exercise behavior

    The finding is not significant after correction for the 28 traits they tested, but it definitely should have been mentioned (it is more significant than other results that were mentioned in the paper). Note that testing as a whole did not increase dietary fat uptake, or decrease exercise, so even if people with increased obesity risk are eating more, it seems that it is being counterbalanced by people with low risk eating less.

  • Barbara myers

    Brave of you to a have your personal genome tested, and brave to publish your results. Actually you should be pleased with what you have learned …not a bad set of results at all and all risks that can be mitigated. Being overweight is actually the real risk factor, but you didn’t need a test for that. Brilliant though if the test has got you thinking about controlling your weigh Good luck with that.

  • incredible post! alas, i told a friend not to worry too much about disease risk in the results…and he came back with a ~50% chance of age related macular degeneration. gotta be careful about that sort of prediction in the future.

  • @Luke

    Sadly that’s not the only significant and arguably important finding that the authors hid in the middle of the supplemental materials. Although the authors conclude that DTC genetic “testing did not result in any measurable short- term changes in psychological health, diet or exercise behavior, or use of screening tests,” subjects whose results showed them at higher risk for Alzheimer’s disease, abdominal aortic aneurysm, and brain aneurysm showed significantly increased test-related distress, and those who tests showed an increased risk of prostate cancer showed increased anxiety at follow up. And that’s my real concern about the industry and the fact that different companies might line up almost perfectly when it comes to analytic validity but we know for a fact* that they don’t when it comes to clinical validity. Being told you’re at risk makes people worry, and the risk information just isn’t consistent.

    *
    Vashlishan Murray, A.B., Carson, M.J., Morris, C.A. & Beckwith, J. Illusions of scientific legitimacy: misrepresented science in the direct-to-consumer genetic-testing marketplace. Trends Genet. 26, 459–461 (2010). http://dx.doi.org/10.1016/j.tig.2010.08.001

    Imai, K., Kricka, L.J. & Fortina, P. Concordance Study of 3 Direct-to-Consumer Genetic-Testing Services. Clin. Chem. 57, 518–521 (2011). http://dx.doi.org/10.1373/clinchem.2010.158220

    Collins, F.S. The Language of Life: DNA and the Revolution in Personalized Medicine. 368 (Harper: 2010).

    Davies, K. The $1,000 Genome: The Revolution in DNA Sequencing and the New Era of Personalized Medicine. 352 (Free Press: 2010).

    Ng, P.C., Murray, S.S., Levy, S. & Venter, J.C. An agenda for personalized medicine. Nature 461, 724–726 (2009). http://dx.doi.org/10.1038/461724a

    Kutz, G. Direct-to-Consumer Genetic Tests: Misleading Test Results are Further Complicated by Deceptive Marketing and Other Questionable Practices. 33 (DIANE Publishing: 2010). http://books.google.com/books?id=aq3Jdd4FqYcC

  • Thanks for sharing this. I have received the results some days ago and it was too much information for me and did not know how to go on… the idea of visiting a GP could be a first step.

  • Justin Loe

    Counterexample:

    A family member tested positive for the A744S Familial Mediterranean Fever mutation. My discussions with a genetic counselor over the phone revealed they had read fewer papers on the subject that I have and were unaware of the possibility that carriers for A744S do exhibit occasional manifestations, albeit rare, of FMF.

    In this case, the genetic counselor knew less than I did about the topic.

  • Interesting overview, good to hear that it has had a positive effect. I still haven’t got round to spending the $162 to get mine done. I would becoming it is from a slightly different point though as I have a significant disease phenotype already (type 1 diabetes) and a strong family history of RA (no surprise there) so I would love to know if I actually have any of the predictive increased risk alleles

  • margaret diamond

    My son chose 23 and Me so that he could get hold of his rs’s and he got the results recently.
    I had had a study done at Pathways Genomics a year or so ago but was not given a complete listing of the SNP’s …. just the SNP’s of the 24 health conditions reported. I think that the ability to tap into ALL one’s SNP’s is an advantage for anyone who has the time to do research. A few of my searches contradicted the alleles that 23 and Me had indicated for a few diseases. I am careful to see that the searches are as current as possible in the medical literature.

  • Ruslan Dorfman

    The statement about bin compound heterozygote for HFE mutations and the note that HFE is a recessive disease does not make much sense: if both mutations are indeed disease-cuasing, and the disease is fully penetrant as CF, then you would have clinically defined HFE. Most likely of one the mutations is just a polymorphism, therefore you do have clinical manifestations of HFE. One of the biggest pitfall of mutation testing is ability to distinguish between disease-causing and polymorphic missense variants in the susceptibility gene. Furthermore, there is a graduate of mutation severity, which frequently ignored, but is the major disease severity factor. See details here for example of mutation severity in CF:
    http://www.ncbi.nlm.nih.gov/pubmed/21175684
    http://www.ncbi.nlm.nih.gov/pubmed/20059485

  • @Ruslan

    Haemochromotosis is a recessive disease with incomplete penetrance, with two mapped relatively high-frequency mutations. The two disease-associated genotypes have penetrances of around 30% and 1%, I have the latter (though about 10% of cases are caused by the latter genotype, due to it’s higher frequency). I haven’t just seen a random mutation in a disease-associated gene and said “Hey I must have haemochromotosis!”, this is a very well characterised disease – we know what are and are not benign HFE polymorphisms.

    As it happens, I do know a bit about genotype-severity relationships in CF. I’ve written a little about the differences between different mutations here:

    http://elainewestwick.blogspot.com/2011/05/finding-out-you-are-cystic-fibrosis.html

    I have also written about common modifiers of CF severity here:

    https://genomesunzipped.org/2011/05/genetics-of-cf-severity-a-survey-of-dtc-customers-and-the-value-of-a-genetic-diagnosis.php

  • Ruslan Dorfman

    Thanks for prompt response – but that is why I posted the comment – incomplete disease penetrance is huge issue that is poorly understood. I believe it deserves a bit more attention and explanation that dietary modification many play significant role in the disease penetrance. You’ve covered it, but I did not feel it was clear enough…

  • @Ruslan

    The post that I linked to, by Kate Morley, covers in some detail the issues of incomplete penetrance, allelic heterogeneity and modifier mutations in haemochromotosis:

    https://genomesunzipped.org/2010/12/digging-deeper-into-my-disease-risk.php

  • Joe Pickrell

    A guy in my building died from the vaccine form of bubonic plague, likely aided by undiagnosed hemochromatosis (link). So don’t get into Y. pestis research!

  • Teddy Devereux

    Thanks for the post. I had a similar unexpected outcome with 23andme. I found that I have a tendency for blood clots (Leiden Factor 5), which by 23andme, I found that I inherited from my mother. I had a blood clot 10 years ago that I assumed was random and due to prolonged sitting during travel. Now that I know I have a risk of having clots, I take precautions of not sitting too long and staying hydrated, and we try to keep my mother active too.
    Both my husband and I also have the same haemochromatosis risk allele, and amazingly our daughter did not get either one.

  • Leigh Jostins

    Thanks for publishing this, Luke. Lots of my friends are interested in getting a test so will be good for them to read this first.

  • Elaine Westwick

    Great post Luke, I’d be waiting to hear your story….Very interesting that you learnt risks for which you had no family history, I wonder if family history would have motivated you quite as much. Well done on maintaining the lifestyle changes.

  • Auntie Jane

    Hello, Luke!
    What an informative and inspiring piece this is, Luke. I’m going to carry on the family research and get tested!

  • James Lick

    Please make sure that you got a complete blood iron panel test. There’s three different tests that are needed to rule out iron overload due to hereditary Hemochromatosis. My own diagnosis of iron overload was delayed 1 1/2 years because only serum iron was tested when I first knew of my risk from a 23andMe test.

    The three tests you need at minimum are: Fasting Serum Iron, Total Iron Binding Capacity, and Serum Ferritin. The first two results are used to calculate Transferrin Saturation Percentage. Divide Serum Iron by TIBC and multiply by 100. Iron overload is determined by Transferrin Saturation 45% or higher and Serum Ferritin over 300 for adult males or 200 for adult females. It is not possible to rule out iron overload without all three tests. If your Transferrin Saturation is over 35 and/or Ferritin over 150 then you should monitor more carefully.

    My Serum Iron has been at the high end of the normal range in each test, but I only found out that I had iron overload when I found out that my Transferrin Saturation and Ferritin were well over the limits.

  • @James

    Thanks for the advice. The tests actually included both both ferritin and transfering saturation, and they both look normal.

  • Thanks for sharing this, Luke Moreover, congrats on the weight loss and getting your blood pressure under control!

Comments are currently closed.

Page optimized by WP Minify WordPress Plugin