Genomic Medicine - Science and Technology Committee Contents

Examination of Witnesses (Questions 480 - 499)


Professor Rory Collins, Professor Andrew Morris, Professor David Porteous and Professor Julian Sampson

  Q480  Chairman: In case we do not all understand the abbreviation.

  Professor Collins: Public Population Project in Genomics (P3G) is about looking at studies of genetics and other risk factors in combination. For example, there is a study of half a million people in China that has recorded similar information to UK Biobank about people's risk factors and exposures. This will be very interesting because in a way the Chinese are where we came from in terms of cholesterol levels—they are much lower; similarly, their body mass index, their weight, is much lower. So we can look at the lower extremes of these exposures which we cannot really study well in Britain.

  Q481  Chairman: What is the size of the study in Japan?

  Professor Collins: In Japan? I am not sure of what studies in Japan you are referring to.

  Q482  Chairman: Perhaps somebody else might help?

  Professor Porteous: David Porteous, University of Edinburgh and Generation Scotland. To the best of my knowledge there are many countries that are starting to think about setting up similar sorts of studies but there is not one which compares with the state that the UK Biobank or the Kadoorie Study in China has reached.

  Q483  Chairman: So what is the difference between Generation Scotland and UK Biobank?

  Professor Porteous: I think there are important differences. There are also important similarities and, more importantly, it is the synergies that we have been looking at. You might say you have just heard that 500,000 participants of the UK Biobank is insufficient, or we would like more is perhaps the way that we would put it, and you might ask what can 50,000 Scots tell us if 500,000 UK citizens cannot? The crucial difference is that rather than asking for single members of the public to come forward to the UK Biobank we are asking families to come forward to work with Generation Scotland. We are doing that because we are primarily interested in the way that genetic risk factors are involved in determining not just disease end points but risk factors that are associated with disease trajectories. So we have talked a little bit about cholesterol levels and that is something that you can measure and measure over a range but the question is, what influence do genes have on setting that level and modifying it on their own and in relation to the environmental exposures? We would like to know whether or not high cholesterol in a mother or father is transmitted and, if so, to what extent to your offspring; and we would like to be able to see the differences in the exposures of the parents in their lifetime, how they compare to the exposures of children, as they grow, through life. So that is why Generation Scotland has those sides to it: our study is to look at the effects as it passes from one generation to another; and it has a trans-generational component to it, anticipating the fact that research we do today is going to be of the most value to the next generation. That is the kind of timeframe that we should be thinking about.

  Q484  Chairman: What is the size of the cohort?

  Professor Porteous: We are aiming to collect 50,000 individuals in one of the three studies that we are doing under Generation Scotland. I will briefly mention the other two, just for clarity. We have, as I say, three studies. One of the studies is simply to collect a very large number of well controlled subjects and that we have done through the Scottish National Blood Transfusion Service. So we have collected 5,000 samples from fully consented blood donors and have from them blood which we can use to do the genetic tests, but also plasma from which we can do bio marker tests. So that will be, if you like, a general population-wide control sample set that anyone who has appropriate ethical approval can come and use as a reference centre. We also have a smaller set of individuals—2,000 we are collecting—where we have not just parental origin but also grandparental origin and here we want to look more at the profile of the Scottish history. So we are selecting those individuals from a number of different discrete geographical locations where we have grandparental history—born in the same region—and we will be able to ask are there differences even within Scotland in terms of the profile of genetic factors and their relationship to health trajectories. That material will also include collections of samples that will allow us to establish cell lines. The importance of that is that although your emphasis is on genomic medicine and I think you are primarily thinking about the impact of DNA-based genetic tests, really we are talking about a whole suite of technologies that allow one to draw lines between the genes, the products of the genes, the R & E, the protein, the metabolites and the cell building blocks of the cell, all the way on to the functioning of tissues and organs and the whole body. That we see as the clinical end of the spectrum. So to fill in those gaps we need to be able to have material that will allow us to ask questions not just about genes but the products of the genes, so this study helps us to do that. But the main study is the so-called Scottish Family Health Study—and I am sure that Andrew will add to this—and the essence there is that we want to study families, we want to study the segregation of genetic markers in those families and see how they correlate with what we call quantitative traits—that is, measureable traits that we can relate to disease processes. By doing so in families we are able to add statistical power that will compensate for the relatively small sample size and in general terms if we get the family collection right we will have more or less equivalent statistical power to a study of ten times that number. The reason we are doing the family-based study in Scotland is because we can do it and because of various other structures within Scotland—the National Health Service for Scotland, the record linkage system, the fact that we have a relatively stable population so that we can follow over the generations makes a family-based study possible. To do so in the breadth of England and Wales is doable but more complex and challenging in terms of being able to do the family history work and to keep track of those individuals because it is a more mobile population. I hope that gives you a feeling for the contrast.

  Q485  Chairman: Thank you, it does. Professor Morris, did you want to add anything about the study?

  Professor Morris: Andrew Morris, University of Dundee and Generation Scotland. I think we are in a rapid transition phase where we are going through a discovery of rare variants with a large effect in terms of individual clinical impact to the discovery of common variants of very modest effect, which are much harder to find and may have much less individual impact, yet can yield great insights into scientific discovery. To do this I think it is clear that we need very, very large studies to be able to have the power and the certainty to tease out the modest clinical impact that many of these genetic variants have. I think that is the basis of these large studies in Generation Scotland and UK Biobank. You may know that it is the 60th anniversary of the NHS this year; it is also the 60th anniversary of study called the Framingham Study in the United States, which in 1948 recruited 5,000 family members, yet the results of Framingham have yielded immense insights into classical phenotypic risk factors, such as blood pressure, cholesterol, smoking, high blood sugar, and we still use data derived from those original participants in clinical practice today. So, for example, if you go to your general practitioner, he may calculate your future risk of cardiovascular disease and that risk calculator is based upon those 5,000 US residents. We use it globally. So I think the origins of these large studies are to be able to adopt the same approach as those original Framingham investigators, but in the genomic era, where we are looking at very small effects in large populations, hence the numbers are so important.

  Q486  Chairman: So the cohort of these studies is likely to yield an understanding of linkage between genetic and common diseases. You are nodding, Professor Porteous.

  Professor Porteous: I agree.

  Q487  Chairman: Professor Sampson, did you want to add anything?

  Professor Sampson: Professor Sampson from the Cardiff University. Another form of biobank of course is to bank samples from individuals who are already affected by specific diseases and this is complementary to the long-term prospective studies that we have already discussed. One such initiative in Wales is the Wales Cancer Bank, which is collecting both tumour tissue and blood for blood DNA extraction from incident cancer cases across Wales. This is a project funded primarily through the Welsh Assembly Government and charities in Wales. This is an ongoing project to create a resource for genetic and allied research into cancer. One difference from the initiatives we have heard about is that the cancer tissues themselves constitute an important element of the biobank and will enable the somatic changes, the changes in the genes that occur that drive tumour genesis to be examined as well as the genes that potentially predispose individuals to cancer through the analysis of the constitutional samples. Of course there are many other similar cohort studies underway in relation to other common diseases that may not be represented in studies such as Biobank or Generation Scotland because of recruitment issues, for instance in relation to psychiatric disorders because those patients are unlikely to take part in these large population-based studies.

  Q488  Chairman: Thank you very much. Could you remind us, Professor Collins, what is the current budget of Biobank UK?

  Professor Collins: UK Biobank is recruiting 500,000 people aged 40 to 69 and the budget for the recruitment phase of Biobank and establishing the systems for looking after the resource, particularly of biological samples but also the data is £62 million, coming primarily from the Medical Research Council and the Wellcome Trust, but with funding also from the Department of Health and from the Scottish Government.

  Q489  Chairman: So that is the initial funding; how much more funding will be required when you get to the stage of trying to find an association of common diseases?

  Professor Collins: It depends how good you want the study to be. The more detailed the phenotyping—that is, the more detailed information you know about the participants in the study—the more informative it will be about the range of diseases. So we are looking at ways in which we can get even more detailed information about a substantial subset of the participants—perhaps 100,000 or so of the participants in the study—which will allow us to look in even greater detail about the relationship between a range of exposures.

  Q490  Chairman: So how much extra funding would be required?

  Professor Collins: The extra funding required to do what we are doing at the moment, which is essentially once recruitment is finished, to continue the follow-up of the participants and to make the resource available for people to use, is probably of the order of a couple of million per year. The way in which the resource has been set up, with a lot of money put into an automated sample storage facility, means that when people need to use the resource they will be able to do so very quickly and at low cost. The money has been spent upfront, if you like, to build an easily accessible resource. So, for researchers who want to use the resource the cost of doing so will be very low, but of course the cost of the actual experiments that they do will depend on the particular type of experiment they want to do and funding of that would obviously be subject to scientific peer review. One comment I would just like to make is that the way in which this study is designed, the primary way in which it will be used is to wait until people develop a particular condition. So in 10 years' time, perhaps 10,000 people will have had a heart attack and you would then take the data and the samples from those 10,000 and from 10,000 similar people within the resource and you would then analyse those samples and those data in exquisite detail. The reason I am stressing that is that we do not measure everything in all half million; we measure the things that really matter in the people who are most informative.

  Q491  Lord Broers: How will findings that you get from biobanks be translated into clinical practice? Are there structured pathways facilitating translation?

  Professor Porteous: I think that this is something which has been, certainly so far as Generation Scotland is concerned, recognised by a number of different groups that are involved in trying to do that translational research, solve the translational research problem. What is the translational research problem? I think there is a recognised gap between the basic science and getting initial understandings about relationships between genes and genomes and health and putting that into practice. We have heard a little bit about how we have done that in the past in terms of the high penetrate single gene disorders but this is a very different matter altogether. So this needs to be a partnership between the researchers, the biobanks, the National Health Service and industry, and industry clearly plays a very important role in this because the one thing that they can do is to take assays and develop targets and from those develop molecules that become drugs and test and have those for clinical use. But to do so effectively is becoming increasingly challenging and so many, many pharma are saying, "Actually the real problem lies in the number of molecules that fail in the clinic." The reasons for those are numerous but one of the major problems is that some of the molecules that look as if they work well in the laboratory or in animal studies just do not transform into things that work well in the clinic. So we have, for example, in Scotland recently had a very large level of investment from WYETH Pharmaceuticals who have come in and said, "We recognise that Scotland has the capacity to help bridge that gap," so we have set up our specific translational medicine research consortium, which is half way placed between pharma and academia, in order to try to better understand why drugs might fail in the first place and to succeed more effectively in taking new drugs through to market. Andrew might want to add a little bit to that in terms of how these relationships might work.

  Professor Morris: I think the definition of translation is important. Cooksey defined two levels of translation. Type one is from bench to bedside and type two is actually getting evidence that is known into clinical practice, from research paper into practice. Sydney Brenner suggested we should have three types of translation in that observations in the population can actually lead to bi-directional translation and lead back to scientific discovery, and I think that biobanks are uniquely placed to contribute to each of those three types of translation.

  Q492  Lord Taverne: I did not quite gather the difference between the first two types. One was to bedside and the second was into clinical practice?

  Professor Morris: We know that there is a lot of existing evidence which is not actually implemented in the clinic, so it is to see something from discovery right through to clinical practice. Type one is from the bench to evidence and type two is from evidence to actual implementation. So it is a two-stage process.

  Q493  Lord Broers: How is this translational research funded? Is it a fraction of the overall biobank funding and, if so, what fraction is that or is it separately funded?

  Professor Morris: I think the phrase translational research has become a major priority in terms of UK healthcare funding priorities and with the creation of OSCHR—I believe you had Sir John Bell, who would have told you about OSCHR. There has been a pooling of resource to try and drive translation and concertina the timeframe from initial scientific discovery to clinical implementation. I think it is currently a focus—whether it is sufficient funding I would not like to say.

  Q494  Lord Taverne: When UK Biobank was started there was some criticism about the large scale of the funding required and also it was thought the whole procedure was rather slow—slow in recruitment and being established. To what extent were these criticisms valid and how far, in so far as they were valid, have they been overcome?

  Professor Collins: As Chief Executive of UK Biobank I think that is directed to me. I would not like to comment on whether or not criticisms were valid; I do not think I am qualified to comment. In terms of the progress of UK Biobank, I think with a budget of £62 million it is appropriate to spend time thinking how you are going to spend it. In particular, to consult very widely, both nationally and internationally, about what questions to ask people about their lifestyle, what measurements to make and what samples to collect and how to collect them in order that analyses can be done that may not yet have been invented. As I mentioned earlier, the idea is that one will go back to these samples in 10, 15 or even (like Richard Doll), 50 years' time and analyse samples using methods of analysis, that have not yet been invented. So a lot of work went into thinking about that; also about how to recruit individuals efficiently into this study in an appropriate way—how to find them, how to invite them, how to assess them at their recruitment visit and, as I mentioned earlier, to store the samples in a way that meant that it would be possible to find a particular individual's sample in 10, 15 years' time. So it took some years before we were ready to start. We then piloted very carefully what we wanted to do and modified that in the light of our pilot experience, and then started the main recruitment in April 2007; we recruited the first 100,000 people by April 2008. We are at about 199,000 as I sit here now, so we will be at 200,000 in October 2008. Our schedule for recruitment was to complete half a million people by the end of 2010 and we are about six months ahead of schedule, and on budget. I think that that is an appropriate way of using such substantial funds. The National Institutes of Health also looked at setting up a similar study in the US. They estimated that it would cost one thousand million dollars and have not been able to do it, partly because of the huge cost but partly because they do not have the luxury of doing a study within a place like the UK where we have a National Health Service, which can really facilitate such studies.

  Q495  Lord Taverne: What about longer term funding? You mentioned that some of the follow-ups would not take a lot of funding, but experiments would be much more expensive; and you also mentioned that the bigger the sample the better—it depends on what detail is required. What do you feel about the long-term funding; do you think it is secure? Are you worried about it? What are your main concerns about long-term funding?

  Professor Collins: As a researcher one is never secure about funding until the grant has been given, but the fact that there has been such a serious investment both scientifically and financially in developing this resource because of the potential value of a resource of this size is encouraging. UK Biobank is, in terms of its size and detailed questions, measurements and the samples, unique; there is no other study on this scale with this level of detail in the world. That is not to say that other studies are not complementary to it, but it is a unique study. So I think that should help to get long-term funding. The amount of long-term funding required to actually maintain the resource—that is, particularly, to follow the health of participants is comparatively small. Even the cost of using the resource in the way I have described, where you only analyse samples and data from the small subset of individuals that are of particular interest for some particular condition means that the cost of doing an experiment are relatively low, and of course you are doing it in five, 10, 15 years' time when the cost of assays that we want to do now are high but as technology improves you can do more; it costs less to do; you are doing it on a focused sample so the costs of doing the experiments are relatively small.

  Q496  Chairman: Are we at the same time investing in these technologies that you will require?

  Professor Collins: For the technologies used to analyse the samples, there is a huge investment going on. Five years ago the idea that you would do what is called a whole genome screen on thousands and thousands of people would have been considered fantasy but now it is routine. The Wellcome Trust is I think now doing several hundred thousand genetic markers on tens of thousands of individuals. Five years ago, as I say, that would have been fantasy. In five years' time we will not even be doing that, we will probably be doing sequencing on tens of thousands of individuals. The costs are falling rapidly in this technology; the ability to get information is increasing greatly. What we have not had until now, until these kinds of studies were being built, is the population resources on which to apply this technology. That is what we are building.

  Q497  Chairman: I am tempted to ask, with all this knowledge that you will gather, can you describe what the scene would look like to, let us say, some young person going to a clinic and asking, "What are my risks?" What would it look like?

  Professor Porteous: I will take the risk of starting to try and answer that. What I would say is it is not like I imagine it will be; it is going to be different from that, but I will try and give you a vision of what might be happening. There is no question that the technologies for analysing gene sequences and very importantly the products of gene sequences, is tumbling and there is a huge technology push to improve these further—there is no question about that. If you simply take, as we have heard, the initial cost of setting up the sequencing of the human genome—3,000 billion dollars for 3,000 million bases of sequence—now we are just at the point where somebody is about to win the prize for sequencing a human genome for $1,000, from start to finish. So the costs are coming down dramatically. But we need, I suggest, to think beyond the genome sequence as the only way to measure risk. The most important way of measuring risk is to ask what is the product of the genes, and that is going to be seen in the proteins that are made by the genes and by the metabolites that are the products of the way that the proteins transform "energy in" to "energy out" and build up the cells, and there are different ways of developing those technologies which are starting to get the same kind of push that the genome technology has enjoyed in the last 20 years or so. That is going to be much more important because that is closer to the actual output, the phenotype, the clinical measures that we are interested in, than simply having everything measured at the level of the gene. So I think that we will see technologies which will replace the kind of glucose dipstick which will now profile all the products that are relevant to the determination of your glucose level or your cholesterol level and so on, and that will be based on evidence that we glean from biobanks and genomic medicine studies.

  Professor Morris: Just to add to that, I think the vision is one of personalised medicine. One could argue that if we give 10 people a drug today, in three people it works, in three it does not work, three people do not take it and one has an adverse event, and that is the current equation for commonly used therapies. So I think pharmacogenetics is part of the vision of what these large studies will deliver, so that we can actually predict therapeutic response reliably based upon a baseline genetic profile. I think in terms of the clinical impact there is pharmacogenetics as well as clinical risk prediction earlier on in the life course of disease progression; so that at an earlier stage we can make predictions of risk so that either lifestyle or therapeutic recommendations can be made earlier on in the life cycle.

  Professor Sampson: If I could answer your question from a slightly different angle and that is I was thinking of the utilisation of these technologies in the clinic in the future. There is a lot that is not known about how that would be implemented in terms of professional understanding and also public understanding and uptake, and so in parallel to these investments in scientific advances it is also very important that social science studies are undertaken and that adequate investment is made in education and public engagement. There are a number of organisations within the genomic medicine community in the UK that are involved in this aspect of the work, including ESRC funded genomics network with centres, including the centre in Cardiff and the centre in Edinburgh, amongst others; and also the gene knowledge parks which were initiated as part of the government's White Paper on genetics, which is still taking a major role in relation to educational initiatives.

  Q498  Lord Colwyn: Do you believe that the translation of the records that you get at Biobank are 100 per cent accurate because we are already hearing about private companies which are analysing the same samples and getting different results?

  Professor Collins: I am sorry, the translation of which records?

  Q499  Lord Colwyn: Of the genomic information leading to susceptibility of disease. Are you happy that you have it right because we are hearing other companies which are taking the same sample from the same person—three different companies—getting different results?

  Professor Collins: Yes, there was a nice example of that in, I think, the Guardian, where a journalist compared the prediction of risk by DECODE, 23andMe and another company and found they varied. Personally I think that the approach of these companies is simplistic. They are based on looking at genetic variants in isolation; they are based on studies which are very small where you can possibly pick up the separate effect of a variant but you do not how it interacts with other things, not just other genetic variants but also environmental or lifestyle factors. So I am actually not surprised at all that they can come up with different results because I think the data on which they are based is weak. That is why we need much bigger resources that will allow us to look at the relationship between a range of different risk factors in order to get a complete picture of a prediction of disease. But I do not think it is only about the prediction of disease. For example, we may well find genetic variants that produce only very small effects on risk, but what that could mean is that we have identified a new pathway for disease. That pathway could then open up the discovery of treatments that would act on that pathway which could be of substantial benefit. An example is the statin drugs which are used in millions of people around the UK and tens of millions around the world, and save tens of thousands of lives each year. The statin drugs developed from observational epidemiology, like UK Biobank, which showed that cholesterol is an important risk factor for disease and from genetic work that showed that there is an abnormality in the receptor—the LDL cholesterol receptor—that influences the amount of bad cholesterol, LDL cholesterol that you have in the blood. The statin drugs interfere with metabolic pathways that can affect the LDL cholesterol receptor. They can make someone who is abnormal genetically normal and lower their cholesterol and lower their risk. It was the combined understanding—I think this is the key thing—of disease process from epidemiology and from genetics that led in that case to drugs that are saving tens of thousands of lives each year.

previous page contents next page

House of Lords home page Parliament home page House of Commons home page search page enquiries index

© Parliamentary copyright 2009