Solving Medical Mysteries Using Sequencing

There is a real “wow” paper out in pre-print at the journal Genetics in Medicine. It is a wonderful example of the application of cutting edge sequencing technology to solve a medical mystery. Even better, the authors also include an auxiliary discussion about the medical and ethical issues surrounding the diagnosis, which raises some interesting issues about the transition from research to clinical sequencing.

The Case

A child manifested severe inflammation of the bowel at 15 months; antibiotics failed to clear it up, and he started to lose weight. Standard treatments seemed to have only sporadic effects, and only severe treatment with immunosuppressants, surgery and full bowel clearing could slow down the disease, which is not a long term solution. No cause could be found; the patient’s active immune system seemed to be acting abnormally, but all tests for the known congenital immune deficiencies came back negative. The doctors could try a full bone-marrow transplant, but without knowing what was causing the disease, and where it was localised, they had no way of knowing if such an extreme intervention would be successful.

Such a severe and early onset disease is likely to be genetic, but testing immune genes at random to find the mutation could take years before it turned anything up. Meanwhile, the child was seriously malnourished, and at times required daily wound care under general anaesthetic. A few years ago this might have been the end of the story.


Getting At The Cause

However, in this instance the doctors decided to order a whole-exome sequence. The 16,000 variants discovered were whittled down in the way that has now become standard, and a single coding variant, predicted to be highly damaging, and not seen in >2000 healthy individuals, was found and clinically validated. The variant is in the X-chromosome gene XIAP, a gene that is known to cause the leukocyte disorder XLP. In retrospect, this make sense, as the authors discovered that previous cases of XIAP-related XLP also showed bowel inflammation. The high risk of death associated with such mutations (most XLP patients don’t live past 10) made the decision to perform a marrow transplant more urgent.

The child had the marrow transplant with few complications, and was eating solid food within 42 days. As XIAP is also expressed in the lining of the gut, there was no guarantee that the transplant would cure the disease, but over the 5 months since the operation the gastrointestinal symptoms have not reoccurred, suggesting that the defect was in the marrow-derived white blood cells and not the gut itself.

The Future of Medical Sequencing

Oddly, this is not the first case of this has happened. A similar paper in 2008 detailed a linkage study of early onset inflammatory bowel disease in an inbred family; the authors found mutations in the receptors of the immune system protein interleukin 10. IL10 is more specific to white blood cells than XIAP, so it is slightly (but only slightly) less surprising that the bone marrow transfer in this case also caused a near-miraculous healing.

Moving beyond inflammatory bowel disease, last year we saw whole-genome sequencing used to diagnose a case of sitosterolemia that was missed by standard tests. This sort of genomic diagnostic technique plays out on a larger scale in projects like DECIPHER, which attempts to identify severe mutations in children with developmental disorders.

Exome sequencing is getting cheaper very fast, often a well-equipped facility can do it for under a thousand pounds. The authors of the Genetics in Medicine paper discuss the cost of the process, and suggest that whole exome sequencing may soon be cheaper than standard iterative testing for immune deficiency patients. Analysing all this data takes resources and expertise, but so does sequencing dozens of candidate genes. 2nd generation sequencing technologies tend not to be clinically certificated (though they are working on it), but that is easily got around by sending off your proposed causal mutation for CLIA-certified capillary sequencing.

The age of medical sequencing is coming fast, and I expect that it will not be long before whole-genome sequencing of severe early onset diseases is routine. It will not always give a clear answer, and will directly suggest a treatment even less often, but as cases like the ones above show, it will be a very powerful tool in the doctor’s diagnostic arsenal.

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

20 Responses to “Solving Medical Mysteries Using Sequencing”


  • Thanks for the summary of this interesting paper.

  • I’ve been a clinical geneticist for 13 years now, and it keeps getting more and more exciting. We will soon be at the stage, as you say, when sequencing will be routine as an almost first line investigation. This poses challenges for clinicians, but it’s a spectacularly interesting problem to have. 7 years ago we were being wowed by BAC arrays. Onward and upward!

  • I think their ELSI approach is as reasonable as one might expect. It doesn’t establish criteria for handling “borderline” results or those that might affect reproductive planning but I realize that one has to stop somewhere if the main purpose is to help a kid.

    6 hours of pre-test counselling is a lot though–I have a hard time seeing that as scalable, even if clinical WGS is reserved for rare cases. At any rate, I imagine that parents who want to help their struggling child will consent to anything they think will work, as they tend to do for other medical procedures involving risk and uncertainty. I am not sure the quantity and range of potential results are as impactful as the prospect of finding a diagnosis and cure, whatever the content of pre-test counseling.

    Something I’ve been wondering is: Are we going to end up with wildly distinct sets of informed consent/results disclosure guidelines for research and clinical sequencing? Symptomatic and asymptomatic individuals? It seems like that would be an impossible mess given the blurring of lines between the two, but I am not sensing any signs of consensus on the issue.

    But perhaps I am just catastrophizing and it will all shake out fine.

  • Well, 6 hours is a lot, but then a lot of these patients will have a heck of a lot more than that in terms of clinical contact before a diagnosis is established. I strongly suspect that as we get a bit more comfy with this approach, and we see how the issues pan out in practice, that will reduce substantially in clinical practice. Sometimes I think we are a wee bit too prissy about the sanctity of our genomes. I agree that incidental important findings (and there will be many, I’m sure) are a poser, but again I think we need to get beyond the notion of the genome being “predictive” (a la GATTACA) and see it more in a “predispositional” light – certainly for adult-onset disorders. This is a pragmatic approach – the technology and cost-benefit factors are coming and there is no stopping them; I guess we need to surf the wave and see where it takes us :-)

  • @Shane McKee

    I think we need to get beyond the notion of the genome being “predictive” (a la GATTACA) and see it more in a “predispositional” light – certainly for adult-onset disorders.

    And sequencing is going to help how in these cases?

  • @Neil: Well, in these adult cases (e.g. the majority of cardiac disease, Alzheimer’s, IDDM, colorectal cancer) it’s probably not going to help at all! At least not in the clinic, for quite a while, until we have large enough datasets to translate sequence into risk, and even then we are still going to be left with the problem that the majority of risk alterations will be from, say 10% in breast cancer up to 15% or down to 5% – whether that is clinically meaningful is debatable. Sure, some major genes will fall out from time to time, e.g. HNPCC or BRCA genes. But in the absence of models where the sequence informs *decisions* that can be taken in the clinic, it’s not going to be terribly useful.
    That is not the case in, say, childhood syndromes, where just getting the diagnosis is often of huge benefit. But people are going to pitch up at clinics with their (self-funded) genomes on USB sticks; we need to have a strategy… Education as to what the genome means and doesn’t mean is pretty important.

  • Shane, do you think that the clinical component of this miraculous and brave new world be sustained in a free market?

  • Well, if they pay for the genomes, they may pay for the consult… I foresee a flood of worried well (or within +/-2SD of well) massing at the portcullis of my battlemented clinic fortress, regardless of whatever disclaimers and riders the private companies put on their reports or in the readme.txt file that accompanies the data. I rather feel I will be telling everyone to give up smoking, exercise more, eat a better diet, and enjoy the rest of their short lives because that unclassified variant in WTF1 looks a bit gnarly.

    Maybe I need to dig a moat.

  • This is a heartening story. Still, it needs to be taken in perspective with the current state of health care.

    I read this story, then about another child with a much more common illness, Strep A:

    (http://www.cnn.com/2011/HEALTH/01/13/emergency.room.ep/)

    It is now routine for hospital emergency rooms to be backed up for hours. Many more children and adults die from this than from extremely rare genetic diseases.

    I’d like to hear more about how genomics will improve public health. Improved broad spectrum antibiotics would be an example. Advancements in the diagnosis and treatment of mental illness would be another.

    The flu?

  • I’d also like to see you put spending on genomic research in perspective with that of AIDS.

    I live in San Francisco. Several years ago I had the privilege of getting to know an AIDS researcher at San Francisco General Hospital. He told me that we could likely beat AIDS with current medicines. The reason for this is that most treated HIV carriers have a very low level of the virus in their systems. They are thus much less likely to transmit HIV. It was his opinion that we had not done nearly enough to diagnose and treat HIV and this is because of the cost.

    Again, in light of the fact that we can’t afford simple public health efforts that could cure major diseases, I’d like to see you put the 1000 genome (US population = 307,006,550 = 0.3 trillion dollars) in perspective.

    I’m letting you off the hook about the fact that the real cost of a full genome diagnosis is probably more than $1,000.

  • But Marnie you could play that game all day. Why not look at the amount of money Americans spend on chocolate or hair colorants? In many ways the knowledge that we’re gaining from the genome is dirt cheap – a lot cheaper than the knowledge we have gained of the internal structure of the proton, for example, and much of the value has been contributed by ordinary families with genetic disorders, working towards making lives better. More than that – genetics is the ultimate study of mechanisms. If you understand *how* diseases occur, you can work out ways to beat ‘em, and HIV is a case in point. Without a lot of knowledge that was hard-won in studying protein interactions and seemingly blue-sky retroviral genetics, we would be much much further back in the fight.

    The fact is we do know how to save millions of lives – rehydration therapy for third world diarrhoea, and vaccinations. If we had the will (or if we put our chocolate money towards Medicins Sans Frontieres, for example), we would solve all sorts of problems. But without knowledge, we will also run across deeper more serious problems that we won’t be able to solve. And don’t forget that having a rare genetic disorder is a *common* problem, affecting approx 10% of families. No-one is talking about 1KGing the whole Western world, but we are getting to an interesting place.

  • Shane,

    I’m not a scietific neophyte so you don’t need to talk down to me.

    I’m not arguing against genomics. Yes, HIV, the flu, better antibiotics are all in the realm of genomics. However, the ability to improve public health will likely have little to do with “personal genomics”.

    I’m not sure why you are talking about Doctors without Borders. I asked you to discuss how personal genomics will improve the treatment of common killers such as strep, the flu, and HIV. You haven’t done that.

    I don’t see that genomic research has been dirt cheap. Take AIDS. Has that been dirt cheap? In the end, we have many terrific medicines, but we can’t afford to deliver them in a way that is broad enough to prevent our own children from having the risk of acquiring HIV. Not cheap at all.

    How about the treatment of opportunistic infections such as MSRA and Strep A. Extremely common. If research is so cheap, why don’t we have more effective antibiotics against these common and increasingly virulent illnesses?

    Schizophrenia? Another disease with a genetic connection. The drugs and treatment for this disease are anything but cheap or very effective. Drugs are only the tip of the iceberg in treatment and they are hardly adequate, available or cheap.

    Personal genomics? It will help a few with rare diseases or the rich.

    Our healthcare system is under a crushing weight of ever escalating costs. In light of this, telling people that personal genomics is cheap, will alleviate health care costs and improve care for the average person is probably less than truthful.

    Really, I’ve been expecting more from the Genome Unzipped team. Some of you seem closed away in the lab, young, and disconnected from the reality of current health care issues.

  • @Marnie

    That is not really fair. There are Genomes Unzipped contributors who work on the genetics of psychiatric illness, and much money and effort has been put into trying to understand the genetic component of these diseases. Psychiatric illness is a good candidate for a personal genomics treatment – diseases such as psychophrenia, bipolar and depression are heterogeneous, and probably represent a range of overlapping etiologies within these symptom-based diagnoses. The response to treatment (be it theropy, drug or lifestyle-change) is also highly variable. Genetics may be able to help us get a hold on these different etiologies, which treatments they do and do not respond to, and hopefully give us genetic, biochemical and behavioural tools to classify and treat them.

    The main reason we don’t have much to say about this sort of study is because, unfortunately, they have not been very successful. For some reason, psychiatric illness has a particularly complex genetic architecture, and simple success stories have been few and far between. Unlike many other diseases, it seems that very little disease variation is due to common variants of at-least moderate effect (which are what the current technology allows us ot study). But this also makes it precisely the field that will benefit most from the fall in whole-genome sequencing technologies, and the tools to analyse them – the falling cost will allow the sequencing of large cohorts to look for rare variation, or common variation of low effect.

  • Hi Marnie,
    I’m sorry if you feel my tone was condescending – that was not the intent, and if it came across that way, I apologise.
    I do think your arguments are based on false premises though. Firstly, it is important to distinguish recreational genomics from personal genomics from public health genomics from rare disease genomics. It is the *latter* where the big gains have been made, and as I said before, I think the payback for the investment has been great, particularly when you compare with other very expensive research projects.

    Secondly, more children die from congenital heart disease than from MRSA, and the solution to MRSA (as a specific example) does not lie in development of antibiotics – we have very effective ways of controlling this pathogen, and pumping money into antibiotics that the bacterium is only going to evolve immunity to anyway seems like a mug’s game to me – unless it is backed up with the very straightforward measures that we already know work. This is not either-or.

    Thirdly, as I mentioned, having a rare disorder is a common problem – I have to look out for my patients. At present, looking after many of these children costs a fortune, and a lot of cash would be saved if we could establish a diagnosis early, and if we could understand the mechanisms.

    So you asked how “personal genomics” will help HIV, Flu & Strep. Who knows? If by “personal genomics” you mean wealthy dilettantes with USB sticks full of sequence, it won’t. If you mean research that pools personal genomic information from lots of (say) flu sufferers and non-flu sufferers to see why some people are more resistant than others, then it could help a great deal (and we have already seen that in HIV). If you mean that doctors can (eventually) use sequence data to target antivirals, then likewise it will help. But “personal genomics” (however you choose to define it) is only one aspect of genomics, and there are multiple drivers at work here, and the cheaper it gets, the more we can do with it.

    Just speaking for myself, I am not *that* young, I am not a member of the GenomesUnzipped team, and I am not closed away in a lab – I deal with real families with rare diseases on a daily basis, consulting, examining, sticking needles in arms, drawing pedigrees, hunting down information. But I am really excited by the potential for developments in this field to deliver real and meaningful benefits for my patients, and I am delighted and honoured to have a lot of young (and old) and enthusiastic lab and clinical colleagues who are making a real difference.

    Sorry that was a bit long, but I agree with you that this is an important discussion – too important to default to an early consensus :-)

  • This might be of interest:

    http://www.biomedcentral.com/1471-2164/9/472

    Polhaus, JR and Cook-Deegan, RM (2008)
    Genomics Research: World Survey of Public Funding
    BMC Genomics, 9:472 doi:10.1186/1471-2164-9-472

    This is extraordinarily bland in its assessment of why genomics gets funded, while more pressing public health needs might not:

    Most countries provide public funding for scientific research, under some variation of a mission to promote or improve the nation’s health. Many countries are investing in genomics as an element of biotechnology, and as a pathway to economic development. The rise of genomics funding results from the priority that governments have placed on such research, which is influenced by policy decisions.

    However, it does have some facts and figures – e.g. Table 5 has that in 2006, the US spent $3.46 per capita on (somewhat narrowly defined) genomics research. Which is a lot of taxes.

    And (with Shane’s caveats that I am neither young nor a member of the GenomeUnzipped team, though I do work in a lab) I did spot one crossover between personal genetics and public health: 23 andMe have some NIH stimulus funding to evaluate whether their web-based questionnaires can be used “improve personalised medicine”.

    http://biomarkercommons.org/biomarker-news/23andme-receives-nih-funding-to-evaluate-web-based-research-on-the-genetics-of-drug-response

    And I just suspect this may be the lever (my emphasis):

    Greater availability of personalized genetic information regarding the efficacy or toxicity of medications could lead to improved patient care and save consumers, insurers and medical institutions billions of dollars per year.

  • Thank you for your thoughts.

    My concern is with setting public expectation regarding whole genome sequencing.

    Shane, we can’t afford to have “a flood of worried well (or within +/-2SD of well) massing at the portcullis of [your] battlemented clinic fortress” and also requesting diagnostic and preventive treatment for illnesses they might not develop until they are in their seventies.

    Improvements in treatment do have real costs. You can google the news to read the latest headlines. It is a wonderful thing that there are improved treatments, but I am concerned that an expectation is being set with the public that will not be able to be met. Medications and treatments will be available, but healthcare plans will likely begin to opt out of covering them, once they realize the cost.

    That’s already happened. A few years ago, there was a rash of articles in women’s magazines about women who had been tested for the BCRA gene and then had a preventive mastectomy and reconstruction. I’m not arguing that this shouldn’t be done, but the insurance companies were quick to react. It has affected far more women than those who are positive for the BRCA gene. Insurance companies have executed subtle but far reaching legal measures to reduce coverage for breast cancer diagnosis and treatment.

    Luke, I appreciate your comments about the difficulty of developing treatment for psychiatric disorders. While psychiatric illnesses are extremely complex, understanding and treating them, while far off, will be one of the true triumphs of genomics, not only for sufferers, but for their families and for society in general.

    Neil, thanks for the links to the 23andme site about efficacy and toxicity of medications and also about the link on genomics spending vs. other areas of health. Very interesting.

    Improved sequencing speed is a necessary piece of the puzzle to be solved, but only one piece of the overall health care picture.

  • Just to put the drug industry and treatment of mental health in in perspective, here’s an article:

    http://www.good.is/post/jared-loughner-schizophrenia-and-mental-health-care-in-america/

    Quoting from the article:

    How good is mental health care in America?

    It isn’t good at all.

    In 2009, the National Alliance on Mental Illness gave the country’s mental health care system an overall grade of D. Arizona, it reported, had an urgent workforce shortage, among other problems.

    A 2008 article in The American Prospect details how state budget cuts, a convoluted bureaucracy, and pharmaceutical industry lobbying have created an ineffective, expensive system.

    On virtually every front, Medicaid and other government agencies, the drug industry, and health-care providers all play a part in a fragmented system that uses taxpayers’ dollars to pay for treatments and drugs that too often don’t work. At the same time, wholesale budget cuts, convoluted reimbursement rules, and conflicting funding agencies have helped create a perfect storm of waste, needless deaths, and ineffectiveness. Tragically, this has diverted scarce resources from proven, recovery-oriented programs such as “supported employment” that promotes a guided return to work, or integrated “dual diagnosis” treatment for the 50 percent or more of seriously mentally ill people who are also substance abusers.

    More and more of Medicaid’s money has been going to expensive new drugs, aggressively promoted by the pharmaceutical industry, that may not be any better than their cheaper alternatives. At the same time, programs to treat mentally ill people with substance abuse problems—the ones most at risk for violent behavior—have been cut.

  • You right!! Think exactly the same

  • Alicia Nossov

    We now know neural tube defects can be largely prevented by folic acid supplements taken before pregnancy. I wonder if bipolar disorder might be largely prevented by fatty acid supplementation ( or other supplementation )before pregnancy in genetically predisposed individuals.
    Has there been enough gene mapping to identify unusual nutritional needs for healthy neurological development?

  • Greetings. I believe there’s an issue with your site loading speed. I hope you can fix it!

Comments are currently closed.

Page optimized by WP Minify WordPress Plugin