Authors: Sarah Earle (University of Oxford, UK)
Genome-wide association studies (GWAS) are a relatively new way to study disease, and this hypothesis-free method can identify associations between genetic variants and traits, shedding light on natural populations.
In this interview we speak with Sarah Earle from the University of Oxford (UK) about her use of GWAS to understand more about Neisseria meningitidis and its transition from an asymptomatic commensal to an invasive pathogen.
First, could you introduce yourself and give a brief summary your career to date?
My background is in biology, I did my D.Phil. in bacterial GWAS, where I worked on the methods and applications with Daniel Wilson in Oxford (UK). I first investigated the potential of these methods using antibiotic resistance as phenotype, and then I moved into investigating virulence in bacteria. That led me to my current position as a postdoc at the Big Data Institute at the University of Oxford where I carried on working on GWAS and bacteria. Currently, I’m interested in the evolution of bacterial pathogens and understanding how bacteria cause disease and become resistant to antibiotics using computational methods.
Could you outline the research you’re presenting here?
Many of the world’s most important bacterial pathogens are common commensals, raising the questions about why these microbes cause disease and whether it is due to bacterial genetic differences within the species. In this work I focused on the exclusively human pathogen N. meningitidis, which is a major cause of septicemia and meningitis worldwide. Although it’s carried asymptomatically in the nasopharynx in large proportion of humans, in rare cases it causes invasive disease and there have been a lot of epidemiological and host factors that have been associated with increased disease risk.
There have also been particular lineages of the species identified that are over-represented in disease. We don’t understand the mechanisms behind that but the fact that these lineages exist suggests that there’s quite likely to be a bacterial genetic basis to increased disease risk. So, that’s what I have been trying to look in to – what is it the bacterial genetic basis of disease?
To investigate this, we carried out a GWAS comparing the genomes of meningococci sampled from carriage and from invasive disease cases – this data is a well-characterized set of data from the Czech Republic in 1993 – and we’re trying to look for differences in the genomes between those two categories.
There are lots of risk factors associated with invasive N. meningitidis disease – how did genomic techniques help you uncover new information?
So GWAS allows us to identify genetic variants important in natural populations as it identifies variants that are actually out there in populations, rather than in laboratory constructs, which can be really useful, but the genetic variants might not necessarily exist out in the wild.
GWAS also enables us to test phenotypes that you can’t necessarily test that easily in the lab, which is particularly important for meningococcus. In addition, we are not confined to testing for particular candidate genes, you can just look at the whole genome and see what’s important – perhaps finding things you might not have been looking for.
Using GWAS we’ve been able to more formally test the lineages that are over-represented in disease and using the genomic data we’ve found a known hypervirulent N. meningitidis lineage, the ST-11 lineage, was significantly associated with disease. We were also able to use genomic techniques to test the proportion of within-sample variability in disease versus carriage and assess what proportion of this is attributed bacterial genetics. And within the data we looked at it was quite a large portion, at approximately 36% attributed to bacteria.
And what variation were you able to identify using GWAS methods and how might those affect N. meningitidis virulence?
So, the presence of capsule is a known virulence factor for N. meningitidis. Using our study we’ve been able to confirm that this is important and we’ve also highlighted variants in genes for capsule production that area associated with virulence. One of the methods to test variation was using kmers – that’s when we take bacterial genomes, de novo assemble them, and chop them into length 31 words or kmers and then test their presence or absence between carriage and disease. By looking at that type of variation, we captured variants in two interesting genes involved in capsule production. The first was in the gene siaD and this is a phase variable gene, meaning that within that gene there is a repetitive region of DNA and the length of that region determines whether that gene is expressed or not.
That’s one of the things we’re able to capture, the second thing is that we were able to capture the promoter region for the capsule polysaccharide modification protein gene, ctrF. It appears in this case the kmers are actually capturing three single nucleotide polymorphisms in the spacer region of the promoter for this gene. So, we think that this is capturing potentially a change of expression of this gene involved in capsule production.
How could your work identifying these genetic determinants of virulence be used for disease prevention or detection purposes?
In terms of prevention, the aim of these types of studies is to try and identify novel virulence factors that could be potentially used in vaccines against N. meningitidis.
In terms of detection, as genome sequencing gets cheaper and is done more in real-time as infections are actually occurring it might be useful for detection and surveillance. If we build up our knowledge base of these genetic variants more, we might be able to identify strains that could either cause disease or cause more severe disease, which would be beneficial. For example, if there is an outbreak we could be able to very quickly identify whether we think this is a very virulent strain, and this could help with ongoing surveillance and monitoring.
Finally, although this work focused on N. meningitidis, do you think this type of investigation might be helpful in understanding other pathogens?
So, there have been a lot of studies recently, really exciting studies, showing these methods for a whole range of different bacteria – looking at really important phenotypes and finding novel genetic elements.
One of the most immediate benefits of this type of analysis could be for antibiotic resistance – if we built up a knowledge base listing what mutations and which elements cause antibiotic resistance in bacteria this could be used to predict which antibiotics will be effective in a specific infection – a step towards a personalized medicine approach. This approach has been used in the NHS for tuberculosis infection – DNA sequencing is used to predict which antibiotics should be used based on resistance profiles – it will be exciting to see that being used more and more for different bacterial species.
You might also like: