Name: Sharon Peacock
Title: Professor of clinical microbiology, Departments of Medicine and Pathology at the University of Cambridge;
Honorary consultant microbiologist at the Health Protection Agency and Cambridge University Hospitals NHS Trust;
Honorary faculty at the Wellcome Trust Sanger Institute
Education: BS in medicine, University of Southampton; MRCP, Royal College of Physicians, London; MS in medical microbiology, University of London; PhD, Oxford University and Open University
Whole-genome sequencing has recently made strides as a tool that can be used in public health surveillance. Most recently, the National Institutes of Health published a study in which they used sequencing in real time during an outbreak to track transmission and make decisions about how to manage the outbreak (CSN 8/20/2012).
Earlier this summer, groups from the UK published studies in the New England Journal of Medicine and BMJ Open, demonstrating the ability of sequencing retrospectively to track MRSA outbreaks in hospitals (CSN 6/20/2012).
Sharon Peacock, an author of the NEJM study, has been making significant gains in bringing whole-genome sequencing into a public health setting.
Under grants from the UK Clinical Research Collaboration and the UK's Health Protection Agency she has been working alongside collaborators, such as Julian Parkhill's team at the Wellcome Trust Sanger Institute and David Aanensen at Imperial College, to develop tools that will enable public health organizations to adopt sequencing for surveillance and diagnostics.
Peacock also published an editorial recently in PLoS Pathogens on the use of sequencing in public health microbiology.
She is currently working within the Cambridge UKCRC consortium to develop databases of pathogen genomes and create interpretation tools for public health organizations, diagnostic laboratories, and health care workers.
Recently, Peacock spoke with Clinical Sequencing News about her work in bringing whole-genome sequencing into clinical microbiology for public health surveillance.
What is the goal of the Cambridge UKCRC consortium?
Our objective is to try and translate whole-genome sequencing into diagnostic and public health microbiology where it's been shown to make a difference either to individual patient care or public health surveillance. At the moment, we're working particularly on [methicillin-resistant Staphylococcus aureus], but we're just in the process of expanding this out to a range of other pathogens that are important for human infection.
Can you describe the work that you're doing with MRSA?
With MRSA, if you look at Europe, the UK, or the US, the majority of infections, transmissions, and outbreaks are caused by a very limited number of bacterial lineages or clones. For example, in the US, you have a problem with USA300; in the UK we have a problem with EMRSA15.
The problem at the moment is that the current technology to type the bacterial isolates doesn't distinguish between one isolate and another. The problem [is that] if you want to try and determine through bacterial investigation whether a strain has passed from one person to another or from one facility to another, if you type the isolates that are associated with that, they all look the same. And so the key thing about whole-genome sequencing is that it gives you a level of discrimination that was previously impossible.
This really started with a paper in Science in 2010 by Harris et al, and that's the key underpinning for our work. What that showed [was that if you] took a bunch of MRSA that looked identical by current typing techniques … when you [did] whole-genome sequencing, there was a huge amount of difference between different strains.
It showed that not only was there clear geographic segregation of isolates — for example, isolates segregated in Asia, or in Europe, or South America — but also, it looked as if you could do sequencing to detect transmission at a hospital level.
So that was very exciting. That really got us going and thinking that whole-genome sequencing really could give us the discrimination to say, 'How is a particular lineage being transmitted?' And the reason we think that's important is unless you understand where things are being transmitted, you have no hope of really controlling the situation, except for blanket control, which is what we do at the moment — everyone washes their hands, etc.
The dream was to really work out if we could detect transmission, ultimately in real time, so that we could bring to bear control measures.
[Now, we've] obtained isolates from across the UK, and we're doing whole-genome sequencing of around 3,000 or so MRSAs so we can really understand the population genetics structure to see if this does map out geographically across the country and also to see if we can track transmission of MRSA from one place to another or within a facility. And the first output of our work appeared in the New England Journal of Medicine in June.
That's our first proof-of-principal that not only can you use whole-genome sequencing to confirm strains in an outbreak and exclude isolates that were not included in the outbreak, but in that paper you'll see that you can do other things. For example you could immediately tell from the sequence what the resistance pattern was.
Now you're in the process of sequencing 3,000 different strains from around the UK?
Yes, and part of that is really building a genetic framework, and that kind of framework is essential for future comparison of isolates. Say, for example, a patient came in with MRSA, either they were infected with or carrying that isolate, and we wanted to try and set it in the context of where it came from. We could only possibly do that if we had an extensive genetic database with which to compare it.
So, isolates from a broad range of hospitals, so we can use that to say, 'What is this most similar to?' In the context of a single hospital, you could have your own strains in a database and you could say, 'What is this closest to? Where did we last see this?'
This database would be accessible from any hospital in the UK?
Yes, exactly. It would form a resource. Somebody could build up their own catalog for their own hospital, against which they compare the next new strain coming along, in which they say, 'Where did this originate from?'
At the national level, the Health Protection Agency would be able to actually undertake more national surveillance using that database.
So it fulfills an individual hospital need but it could also build a much more extensive picture of MRSA at the national level, or even international level in time.
What is the time frame for building this database?
The strains are largely sequenced. Now we're doing bioinformatics analysis. And I think it's important to stress that the next impediment that we need to overcome is to develop interpretation tools that can be used by people who don't have expertise in bioinformatics and that provides an output in a format that can be understood by health care workers and public health workers and laboratory technicians.
So we've certainly sequenced a large number of isolates to provide a genetic database, but our next job is to work on a technological solution. So, for example, somebody can actually include their own sequencing to the database, and an automated interface would provide meaningful information back to the clinician. We're working with a group at Imperial College who are undertaking some of that development work at the moment.
Are you doing this for other pathogens?
Yes. We are intending to try and build databases for most, well, all of the major human pathogens. We're focusing on bacteria at the moment. So we're currently building a database for vancomycin-resistant Streptococci. This is a bacterium that tends to infect people that are very immunocompromised, for example transplantation patients. It also spreads within the hospital setting. It's a problem because it's resistant to vancomycin, which is the treatment of choice.
We're building a database for vancomycin-resistant Streptococci, for gram-negative organisms, and for other important human pathogens that are responsible for outbreaks in a hospital setting or outbreaks in the community, [such as] foodborne pathogens.
In the editorial in PLoS Pathogens, you mention a couple of different possibilities for implementing microbial whole-genome sequencing in a hospital setting, such as having a centralized reference lab be in charge of doing all the sequencing versus hospitals themselves. Do you think one scenario is more likely than the other?
I think the scenario for that will change over time. I think that initially, because this technology is a new one, I think it's likely to be taken up by a limited number of centers. Not necessarily centrally. There may be some central activity, for example the Health Protection Agency may wish to undertake sequencing. But I predict there will be a limited number of fairly large centers that will be undertaking this technology and using a central database for interpretation of that data.
I think ultimately the technology will spread to other places. I think that in the first instance, it won't be used in every hospital, but 10 years down the line, I suspect it will be fairly familiar, as PCR is today. Every hospital does PCR these days, but one wouldn't expect every hospital to do whole-genome sequencing until such time as it's been proven, and is familiar and the technology is very easy to use.
You've mentioned interpretation is one major hurdle. What are the other hurdles?
I think there are two. The first is an interpretation tool or tools and an interface that people can use. I think we also have to address the issue of data storage, because this will actually generate a large amount of data and that sort of data would be quite challenging for regional hospitals, or even a central database, to collect. So I think that the issue of data storage is going to have to be addressed.
And I think the third thing is working out where these applications make a difference. So economic analyses to work out where we should be applying this and where we shouldn't be applying it because it actually doesn't bring any benefit to patients or populations. If we show that certain applications are cost effective, then they're likely to come into use. But if we show that something is of academic interest but has no real value in the clinic, I think that it will have a very limited shelf life.
In terms of the sequencing technology itself, do you think the technologies currently available are suitable?
I think that the field is changing so unbelievably fast. Just a very short time ago there were no benchtop sequencers. The introduction of rapid turnaround benchtop sequencers does mean that the technology has arrived. We can actually do it in the clinic. But I think that cost still does need to come down. And this is limited by the need to undertake fairly complex bioinformatics analyses. So we can do it, but I think it remains very complicated at the moment.
I think that when interpretation tools become available then other people could start to use it outside of the academic arena. The other thing is I think we haven't seen the end of technological improvements. There's new technology coming along, for example, Oxford Nanopore, which looks very promising. I think that it's a bit of an express train, and it's going to keep moving at the current pace until, I don't know where it ends. There are newer technologies already on the horizon that look very promising and might supersede the current ones. But the current technology will allow us to do sequencing in a clinical setting provided we've got the backup to do the bioinformatic interpretation.
How soon do you think whole-genome sequencing in a diagnostic, public health setting will be used?
That's a million-dollar question. I think it will depend on the application. For example, sequencing has already been used retrospectively to investigate the E. coli outbreak that happened in Germany (IS 6/7/2011). I wouldn't be at all surprised if during the next big outbreak, sequencing is used intensively to try to understand it and control it. I would anticipate that people are on the cusp of using genome sequencing for particular, very high-profile problems. I think that we're very close to that.
The HPA had the ability to do whole-genome sequencing if an outbreak had occurred during the Olympics (CSN 8/8/2012).
If we look into the future, I think the timing will be dependent on interpretation tools, but I would imagine that we would start to see the use of technology that has shown to be effective in the next two to three years. Perhaps shorter than that, that's pessimistic. I'm sure people at the HPA will be using sequencing to undertake outbreak investigations sooner than that.
What sequencing technology are you using?
The reference database we're building is using the HiSeq, and the work that we're doing in the clinic is using the MiSeq. We have several studies that we've completed that are currently unpublished that use the MiSeq. We've completed studies on MRSA and TB that are under review.
Are there any regulatory hurdles or issues with privacy that would make implementing this in a clinical setting difficult?
I think that when you're dealing with patient information, ethics and data protection is always going to be an absolute priority. But if hospitals are performing diagnostic tests, then they're already very familiar with those data protection issues.
If you were submitting information to a database then it would be completely anonymized and so nobody could actually tell where that came from.
I think that one thing that may be an issue in the future is that if we sequence directly from a clinical specimen that contains human DNA, then sequencing the entire human specimen would actually generate sequence of the human DNA.
One of the areas we have to be very aware of is that we mustn't fall into the trap of sequencing human DNA inadvertently. There's methodology available to make sure that's filtered out. In terms of confidentiality, I think that people who look after patients have the mechanisms in place to make sure we don't break the ethical code or privacy code, but we do just have to bear that in mind.
But [in the near term] we're going to be sequencing from bacterial colonies, so that won't be contaminated with human DNA.