Derek Morris, Senior Lecturer in Biomedical Science at the University of Galway, joins us to discuss the genetics behind psychiatric disorders, including why they are currently so difficult to treat effectively and his recent paper looking into the genomic regions linked to schizophrenia risk.
Please note the transcript has been edited for brevity and clarity.
FLG: Hello, and welcome to the latest “A Spotlight On” interview. Today I’m joined by Derek Morris, Senior Lecturer in Biomedical Science at the University of Galway, who is going to talk to us about risk genes for psychiatric disorders. Without further ado, Derek, could you introduce yourself and tell us a little bit about what you do?
Derek: Hi there Lauren, nice to speak to you today. My name is Dr. Derek Morris, I’m a senior lecturer in biomedical science at the University of Galway. My background is in science all the way – I did a degree in biotechnology here in Galway before doing a PhD in molecular genetics at Cardiff University. And that’s really where I got into the area of psychiatric genetics. I followed that up by taking on a postdoctoral position in Trinity College, Dublin. I went on to take a position as a visiting research fellow at the Broad Institute of MIT and Harvard University before becoming a lecturer at Trinity College Dublin, and then moving to a lectureship here in the University of Galway at the end of 2013.
My research interests are in psychiatric genetics, and that’s about understanding or identifying risk genes for common psychiatric disorders, such as schizophrenia, bipolar, and major depressive disorder. The main goal of what I do centres around a central question. And one way to look at that is to maybe ask, what’s the difference between a disorder and a disease? So, one answer is that a disorder is an illness that disrupts normal functioning, whereas a disease is an illness with a known biology. My research is really based on trying to turn psychiatric disorders, which is the term we use, into psychiatric diseases, which is a term we don’t really use. And we don’t use that term because we don’t really understand the biology of these illnesses. But importantly, these disorders have high heritability, which means that genetics plays an important role in causing these illnesses.
So my research is focused on trying to exploit that – trying to find the risk genes for, for example, schizophrenia. And if we can find genes, we can understand what proteins are involved – what biological processes are involved – and from there, try and understand the biology of the disease.
FLG: It’s a really interesting area, and I’m really excited to get into some of these questions. First, I’d love to know, what inspired you to go into this specific area of research?
Derek: Yeah, good question. It stemmed really from my interest in biology, even in secondary school. Specifically in genetics. I remember watching a program – a BBC Two documentary – that described the size of the human genome and how many bases may be in it in terms of the As, Cs, Gs and Ts. I also had an interest in mathematics and an interest in biology. And I think when the two of them are combined, you can have an interest in genetics, because it can be quite kind of computational and mathematical as well as biological. When it came to getting into my postgraduate study, I was interested in doing a PhD abroad at the time – there were different offerings, but the one in Cardiff was quite attractive and it happened to be in the area of psychiatric genetics. So, I hadn’t specifically picked psychiatry up to that point, but since my time in that lab, it has established itself as a world leading laboratory in terms of psychiatric genetics. I was very fortunate to get some really good training there. And I’ve ended up kind of remaining in that field over the last 20 years.
FLG: Can you tell us a little bit about the Psychiatric Genomics Consortium?
Derek: Yeah, so the Psychiatric Genomics Consortium was formed in around the late noughties or into kind of 2010/2011. And it was off the back of an initial wave of what we call genome wide association studies or GWAS. So GWAS occurred in the mid noughties and they were the first application of the SNP-ChIP technologies that led us to be able to analyse large numbers of genetic variants across the genome, and to assay them in both patient samples and in controls. And then to compare the data from patients and controls to try and find genetic variants with different frequencies between the two samples, which would point to what we call a genetic association, which will help identify risk genes for the disorders.
Initially, we formed what was called the International Schizophrenia Consortium in late 2006. And that brought together about eight institutions, mostly from Western Europe and from North America, who had collections of schizophrenia patient samples and controls. Prior to this, there had been very little data sharing or sharing of samples between sites, and individual institutions and laboratories were often quite protective of their data and weren’t too keen to share it. But what the initial GWAS studies told us was that if they were to be effective, we needed very large samples of patients and controls. The reason for that is that it turns out that common genetic variants, which may which contribute to the risk of these disorders, have relatively low effects or low odds ratios, which means you really need large samples to be able to detect true associations.
So that really forced laboratories and institutions to come together. And as part of the international schizophrenia consortium, we published two major papers in Nature in 2008 and 2009, looking at both rare copy number variants, and common SNPs, and these are the some of the first major genomic studies of schizophrenia. Off the back of the International Schizophrenia Consortium, there was a recognition in the field that the consortium needed to open to more institutions, and also to extend to the study of other disorders. It now comprises hundreds, if not 1000s, of investigators, who pool and share samples and data for the purposes of gene discovery. And it has extended well beyond schizophrenia to now include other common disorders like ADHD, autism spectrum disorder, bipolar disorder, major depressive disorder, but also disorders like Tourette’s Syndrome, anorexia, and others. So it’s huge and varied. The website actually is a great resource for both the public, for scientists and for clinicians, if you want to learn more about the Psychiatric Genomics Consortium.
FLG: Yeah, that’s fantastic. You’ve sort of mentioned it there, but I think it’d be interesting to delve in a little bit more – why is it so important to have this sort of collaborative work going on, especially in this field?
Derek: Yeah, so before GWAS and before the genomic revolution – which came off the back of the first draft of the Human Genome Project – about 20 years or so ago, there were a lot of genetic investigations into psychiatric disorders, just like other common disorders like cancer, cardiovascular disease, asthma and respiratory disorders – you name it. And these largely had been what were called candidate gene studies – they largely focused on individual genes, and individual laboratories with maybe a few 100 patients and a few 100, controls. Now, occasionally, they would find significant associations. But invariably, these associations didn’t replicate when they were studied in independent samples, which is a key factor for genetic associations – you need independent evidence to support your initial statistical findings, in case your initial findings were false positive results.
There was also the recognition that we needed bigger samples to be able to detect initial associations and also to be able to replicate them independently. GWAS chips were not cheap, and all of this research was going to cost huge amounts of money. Therefore, funders were going to stump up for that. But in response to that the funders, for example the Wellcome Trust, identified that they would only fund this type of work if large groupings came together such that sample sizes of sufficient size were available. There was also a recognition that whatever data was generated, it needed to be made available down the line for other researchers to use it and reuse it to try and get as much out of that data.
So, again, primarily, the main reason these large sample sizes are needed is because of the individual genetic risk variants, for example single nucleotide polymorphisms – these individual changes in our DNA code (we’ve got millions of these in our genome). And for the common ones that contribute to risk of different traits and affect different phenotypes, they only have a small effect, so there’s only a small amount of increased risk associated with carrying a particular allele at a particular schizophrenia risk SNP. What that means is that when we compare patient samples to controls, the frequency difference between those two samples in terms of allele frequency is relatively small – in the order of 1 to 3% – so to detect a statistically significant association, where the frequency difference is so small, you need very large samples to have accurate measures of allele frequency in your patients and in your controls.
These are the main reasons why large samples need to be collected. And what we’ve learned in the 15 years since the initial GWAS were undertaken is that when our samples get bigger, there’s an immediate payoff in terms of the new risk genes that we can identify. The more samples you have, the more power you have to detect these low effect associations. And that has driven gene discovery in psychiatric disorders, but also in lots of other common disorders as well.
FLG: Yeah, that makes a lot of sense. I guess this kind of links into that as well then, but certain neurodevelopmental disorders are polygenic. Could you explain what that actually means and how it can impact how we study and treat these disorders?
Derek: Yeah, so polygenic simply means many genes. And again, before GWAS and even in the early GWAS days there was a lot of debate about what was the genetic architecture of these disorders – how many genes might contribute? So we knew from older linkage studies, going back to the 80s and 90s, was that it wasn’t a single gene disorder, like cystic fibrosis or Huntington’s disease. And actually, the scale of its pathogenicity has gone way beyond what we might have expected. In current studies, even though we’re finding hundreds of risk genes for schizophrenia, the truth is that they may be hundreds among 1000s that are actually in our genome. So there’s many, many variants that are contributing to risk. And again, because of these low effect sizes, it means that we need very large samples to be able to detect them in terms of what it means in terms of treating the disorders.
To treat the disorders, we need a better understanding of the biology. And again, we’re trying to drive these new biological understandings via genetic discoveries. So if we find the genes involved, we can understand the proteins that are contributing, we can understand the biological processes that are involved, the cells and the tissues that are involved, and from there, hopefully, develop new drug targets and be able to improve treatment. In terms of treating these disorders, the new genetic studies haven’t yet had an impact. But it doesn’t mean that they’ve not been successful yet. It’s just a long road to translate genetic findings into new treatments.
FLG: Could you provide us with a bit of an overview of which treatments are currently available for schizophrenia and also why our limited knowledge of the underlying biology is impacting the effectiveness of these treatments?
Derek: Current anti-psychotic medications that might be used for schizophrenia were discovered by accident nearly 60 years ago. Since then, there’s been few major discoveries in terms of new medications that have become available for patients, which is pretty shocking. If you compare that to other common illnesses and disorders out there, whether it’s cardiovascular disease, or cancer treatment, etc, there’s been revolutions in terms of new treatments that are available to treat disease and prevent disease. But sadly, for psychiatric diseases, there’s been relatively few breakthroughs. And even for the medications that we do have, they’re only partially effective. They’re effective for some individuals and not others, they’re effective for some symptoms of the disorder and not others. Typically, they can be effective for treating positive symptoms, such as hallucinations and delusions, but they’re not so effective at treating negative symptoms and cognitive deficits. And it’s those aspects of the disorder that are often the most disabling and prevent individuals that are affected from leading normal lives in the community. This results in what we call functional disability. So we really need to get a handle on the biology of the illness so that we can identify new drug targets. And again, that’s where our genetic research comes in. If we can find the genes, we can hopefully find the targets that would encourage new drug development.
FLG: Traditionally, why has it been so tricky to understand the underlying biology? Does it sort of just come back to lack of genetic knowledge?
Derek: Yeah, so it’s lack of genetic knowledge in terms of cause. I mean, as a brain disorder – the brain is pretty inaccessible. When you compare this to a disorder that might affect your other tissues, we can’t get tissue samples to undertake a biopsy, etc. So we’re relying on collecting clinical data on patients, collecting behavioural data on patients, we can do neuro imaging studies to look at the structure and the function of the brain. And we can combine that with genetic studies. The point is we have to find alternative routes to try and understand the pathology of what’s happening and identify the molecular mechanisms of disease. And given the, again, the inaccessibility of the brain, and it’s also incredible complexity – it can prove very challenging. But, again, genetics is a route to solving that problem, and that’s what we’re trying to exploit.
FLG: Could you tell us a little bit more about the paper that you published earlier this year, where you and your collaborators identified genomic regions affected in schizophrenia risk?
Derek: Yeah, so this was the latest paper by the Psychiatric Genomics Consortium. And what it represented was – since the previous kind of “major” paper in 2014 and another major one by my colleagues in 2018 – further efforts to collect larger samples and undertake bigger GWAS to identify more risk loci, as well as combining it with datasets that have been collected from other regions – primarily from Southeast Asia, but there were also samples from Africa included in this study this time round. And this paper was in line with what I said earlier – that as the samples get bigger, the more loci we’re able to identify. So it succeeded in doing that, and finding between 200-300 risk loci for schizophrenia. Based on that, then there was various analysis to look at the tissues, and particularly the cell types that were affected, and there was a big drive towards trying to identify causal variants.
A big challenge in genetic studies is that when we do a GWAS and we find a region of the genome where common variation increases risk of illness, we need to then try and figure out which of the common variants present at a region are likely to affect gene function and how. Based on that, we then form a hypothesis about how those genes may be contributing to disease. So there were advances in terms of fine mapping techniques. There was also overlap with other studies, and some of the headline results relayed that a lot of the risk genes that were identified (when we look at their function) map to neurons, and specifically, are involved in synaptic function. It turns out a lot of the genes that are risk genes for schizophrenia encode proteins that function at the synapse. So it really gives us an area to focus on in terms of biological advances.
FLG: That’s a great overview. So you sort of mentioned there how some of these areas map to the neurons. But does this actually have an effect on schizophrenia symptoms?
Derek: We don’t know the answer to that yet. So obviously, it does suggest that the molecular mechanisms of disease are certainly at play with the neurons and again, specifically with the synapses, the connection point between neurons. But how exactly that’s resulting in the symptoms that patients experience, we can’t yet answer. But it’s part of our follow up analysis, where we can specifically look at genes with synaptic function, and start to correlate that and compare that with different symptom profiles and patients, because different patients present with different symptoms. And hopefully, that will start to generate some answers to the question that you just asked. But as of yet, we can’t make that firm causal link between the synaptic bio biology and the symptoms that patients experience.
FLG: You might not be able to tell us in great detail about this, but what sort of areas are you looking to work on in the future?
Derek: What these big GWAS studies do is they identify hundreds of risk genes across the genome that are contributing somehow to disease risk. The next big challenge is trying to take those results and building on them to , for example, how individual genetic variants affect gene function, be more certain about the exact genes that are involved, look to see which of those genes may function together and where they may function, what they may do, and understand some of the biology that’s there.
In our laboratory, we’ve taken a particular interest in genes that function as transcription factors during development. What that means is that there are genes that regulate the expression of other genes that control which genes are turned on and turned off during brain development. Because these genes are often associated by GWAS and rare mutations in these genes often cause severe developmental disorders and intellectual disability. So what we try and do is explore data which gives us a map of the genes regulated by these transcription factor genes and to look within these particular networks to see if the first enrich for genetic risk in schizophrenia and if that’s the case, maybe what cell types are being affected, and at what point in development the effect is occurring So again, these GWAS are really a foundation for lots of different Advanced Biology and computational science to try and understand the biology that’s involved.
FLG: That’s really interesting. So are you looking more into sort of transcriptome analysis now in your research?
Derek: We do a different level. So, there’s transcriptome analysis available in psychiatric research, there’s some that’s available on post mortem brain samples from patients and controls. But another big area to look at is to use single-cell analysis. And we have several projects in this space.
So single cell analysis, or single cell RNA sequencing analysis particularly, is based on being able to look at the transcriptome or look at gene expression patterns in individual cells. Previously, gene expression analysis would often have relied on bulk tissue. So that may be for a specific region of the brain, such as the prefrontal cortex. But when you get a sample like that, and you perform gene expression analysis, what you’re getting is expression data mixed together from genes that are expressed in lots of different cell types. What single cell analysis allows us to do is to separate out the individual cell types and look at the expression patterns in each cell type. For example, in different types of neurons compared to oligodendrocytes, compared to glial cells, etc. And what that lets us do then is that when we get GWAS results, we can look to see which particular cell types those genes have high expression in. And what that helps us identify then is the individual cell types where the molecular mechanisms of schizophrenia may be at play.
That’s also important in the context of drug development, because I’ve been to several presentations by researchers from the pharmaceutical industry, and they will always highlight the importance of knowing what cell type needs to be targeted with the drug. So hopefully, if the genetic research and the single cell analysis can proceed with a better idea of the important cell types from schizophrenia, that gives the pharmaceutical companies a narrower target for their drug development activities.
FLG: Is that something that hopefully will progress a bit more? That we’ll see a bit more investment from pharmaceutical companies in these treatments?
Derek: Yes, hopefully. As I mentioned, for pharmaceutical companies, identifying the cell types is a key aspect that needs to be addressed. There’s lots of exciting new research coming out that is looking to try and compare schizophrenia risk genes and their expression patterns in cell types, not only different cell types from different brain regions, but also looking at different stages of development. Because, you know, we understand that schizophrenia is often described as a neurodevelopmental disorder, which means that in part schizophrenia is due to abnormal your development, even though individuals don’t present with symptoms until their late teens or early 20s. But it’s likely to be in part a consequence of the brain not forming properly or correctly, whether that’s sufficient synaptic connections or whatever it might be. So as well as looking at cell types, we also need to think about which cell types are might be affected at which time points in development.
There’s lots of effort to try and do that. It’s supported by single-cell analysis that’s available on human brain samples. But there’s also big resources that are made available, like studying the mouse as a model organism. And even though mice and humans are different, obviously, there’s lots of similarities in terms of the biology of brain function. So those big single cell resources from mice are helping us identify what cell types might be important for schizophrenia.
FLG: That’s really interesting. I also wanted to ask you about another one of your recent papers which focused on maternal immune activation. This isn’t something that I’ve come across before. Could you explain what this means and how it impacts psychiatric disorders like major depressive disorder?
Derek: Yeah, so maternal immune activation (MIA) is where a pregnant mother experiences for example, a bacterial or a viral infection, and that results in an immune response. And that immune response can have an effect on the developing foetus and in particular can affect brain development. What we know from epidemiological studies is that this represents a risk factor, not just for major depressive disorder, but for many neurodevelopmental and psychiatric disorders – schizophrenia and autism especially.
This is an example of an environmental risk factor, we have mostly been talking about the genetic risk factors for these disorders. But of course, these disorders are caused by a combination of genes and the environment – a combination of nature and nurture. And MIA is an example of an environmental risk factor. This is something that we’ve been able to study, like other groups, in an animal model. And that’s where we can try and understand the effect of an immune response, we can look to see what the effect of that is on the brain function of offspring and try and relate that back to human studies.
FLG: Can you tell us a little bit more about what gene-environment interactions are and why it’s so important to study these?
Derek: Yeah, so gene-environment interactions occur when genetic risk factors and environmental risk factors combine together to increase risk of illness above and beyond what they would in isolation – in an additive manner. So each might individually contribute to risk. But a gene-environment interaction may be when an individual is subjected to both, then perhaps it has a multiplicative effect on their increase in risk. They’re important to study because it will help us understand how gene function may be perturbed by environmental influences, it helps us understand how the environmental factors may be functioning.
However, they are difficult to get a handle on and they’re difficult to study. So you need genetic data and environmental data. Genetic data is easier to collect because an individual’s inherited genome and the common risk alleles that we have in our genome, those that are inherited from our parents, are largely stable and constant over lifetime. So one genetic analysis using a SNP-ChIP as part of your GWAS analysis, will, for an individual, catalogue the different variants they have in their genome over the course of their lifetime. On the ither hand, environmental data is harder to collect, because it varies over a lifetime.
There’s a huge range of different factors that you can collect data on, whether it’s maternal immune activation, whether it’s major psychological stress, stress at different time points in life, whether it’s use of cannabis, whether it’s other traumas, etc. So it’s best to try and collect environmental data prospectively using longitudinal studies, which means that you try and recruit individuals into a study and collect data on them at different points in their lifetime. This is both expensive and time consuming, but there are large biobank collections underway that collect such data. And over time, hopefully, we’ll be able to better form gene-environment interactions, but at present, it is hard to pinpoint to any specific gene-environment interactions where we can be confident they are true associations and would be replicated in independent samples.
FLG: Are there any sort of developments or new technologies in this space that could help with collecting that sort of environmental data? Or is it just a case of having to wait it out and do the longitudinal studies?
Derek: There are various advances that are useful. I mean, part of it can just be effectively collecting health data with patient consent over time. And that is facilitated more when hospital systems have electronic health records. And again, if those individuals consent to be part of large studies that combine their health data with their genetic data, that’s useful.
We’ve obviously kind of got wearables and modern technologies – our watches, our phones, etc, which do collect data on our behaviour and some aspects of our physiology. And now in the areas that we’re interested in, whether it’s psychological data or neurocognitive data, that data isn’t as easy to collect individually over time in large samples, but certainly for other types of studies, such as in cardiovascular disease, a lot of the data that’s routinely collected by our other technologies can be used, again, with patient consent as part of major investigations of gene-environment interactions.
FLG: As the price of whole genome sequencing continues to decrease, what will this mean for the field?
Derek: Cheaper whole genome sequencing means that it’s obviously more affordable. And it means, much like the SNP-ChIPs, that we can do genome sequencing on large numbers of samples. So why would we do that? Most of the major studies undertaken to date in schizophrenia have focused on common SNPs. However, there are other types of variants that may contribute to schizophrenia risk, and they are rare variants. And that’s where the change in the DNA code is relatively unusual and very infrequent in a proportion of the population. So when we undertake SNP-ChIP analysis, as part of GWAS, what we’re doing is just asking or measuring the common variation across the genome, which doesn’t tell us much about the rare variation – with the exception of very large deletions and duplications called copy number variants.
If we want to find more of these rare variants, we need to be able to sequence chromosomes essentially, from end to end in each individual patient and use that to identify rare variants that may increase the risk of illness. In tandem with the earlier paper I mentioned, another paper was published based on what was called the scheme of study. And this was a study that used a combination of exome sequencing and whole genome sequencing to look at rare variation in schizophrenia.
So just briefly, exome sequencing is where we sequence just the exons of the genome as opposed to the full genome. So it’s only about 3% of the genome. And for a long time, it was a lot cheaper than whole genome sequencing, but that’s slowly changing. And the reason we study the exome is that it’s easier to interpret rare variants in exons, because we can predict what effects they are likely to have on protein structure and function, etc. There’s a whole other area of research to undertake, which is looking at rare variants outside of protein coding regions, they’re also important, but it’s going to take a lot more work to be able to interpret those.
The outcome of the schema study was to identify 10 schizophrenia risk genes across the genome, where there was an excess of rare mutations in patients compared to controls. And again, this gives us valuable insights into the life and possible biology of the illness. Importantly, there was some convergence in terms of the biology in that compared to the GWAS studies, some of the same genes were identified, and some of the same biology was also identified. So that’s very encouraging. And, of course, the identification of genes that have rare mutations, immediately helps us identify what are likely to be the causal variants for those individuals that carry those mutations. And therefore, we can start to model them a little bit easier in cellular models.
Going forward, I would say that the gene discovery process for schizophrenia will continue with both of these arms and the way I think GWAS will continue is it expand into populations of different ancestry around the world. And they will look to identify more low risk variants but there’s also going to be big push to do more sequencing studies, and to use that to drive rare variant discovery. And the two in combination, hopefully, will give us greater insights into the genetic basis of the illness.
FLG: You identified around 10 genes I think you said – will future work focus on looking into those in a bit more detail?
Derek: Yeah. So again, in comparison, in GWAS studies, where we identify a risk gene or risk locus, what we’re identifying is a common variant and other variants in high linkage disequilibrium with it, which means that they’re inherited in a correlated manner. But often it can be difficult to move from the genetic variant to the risk gene, and that’s because of this high linkage disequilibrium. It makes it hard to figure out, first of all, which gene is affected, and then secondly, the DNA variants, you know, even if we know which gene it does affect how does it affect the gene? Does it change its expression does it change its function etc. And so that makes it challenging then to be specific about which genes to bring forward for functional studies.
When we do rare variant analysis, like the schema study that identified just 10 genes as opposed to 200 plus from the GWAS study, for those 10 genes, what we have is a catalogue of specific mutations that affect the gene function and affect the protein function. Now, often these are what are called loss of function mutations. So it means that there’s a change in the DNA code that stops the protein from functioning, or else there may be missense variants, which means that they change the amino acid in the protein, but the consequence is that it functionally impacts on protein function. So that gives us a more specific handle on the causal variant at that gene. And therefore, again, if we want to model that in a cell, or in maybe in a brain organoid, or some other type of model to investigate function, it means we’ve got a more specific idea of exactly what genetic lesion or what genetic mutation we need to study, what particular gene is involved, what the effect is on the protein, and trying to again understand what the biological consequences of that are on cellular function.
FLG: One thing I think you mentioned in that or something that comes up a lot with genome wide association studies is around the lack of diversity. Could you tell us a bit more about how this lack of diversity can impact on genetic findings around psychiatric disorders?
Derek: The issue around lack of diversity is that GWAS studies of the last 15 years have predominantly been based on Caucasian samples of European ancestry collected either in Europe, or maybe collected in North America. There’s also been collections elsewhere, such as in Australia, etc. There have been collections of more diverse samples, for example, African American samples collected in the US. And there have been more recently, some large collections of samples, for example, in East Asia. However, there’s many parts of the world where that are underrepresented or not represented at all in genetic studies.
Now, the PGC has made progress on this, but there’s a lot still to do, the progress that has been made is that there has been a large, East Asian GWAS undertaken of schizophrenia, that was published by Lamb et al in 2019. And importantly, in that, when they compared the GWAS results from the East Asian samples to the European samples, that indicates that the genetic basis of schizophrenia and its biology was largely shared across the population. So there are different risk genes in one population of one ancestry compared to another. However, when they attempted to transfer the polygenic risk scores between populations or between ancestries that wasn’t really effective. And what that highlighted was the importance of sufficient samples of different ancestral groups to ensure the that we can generalize across populations when it comes to polygenic analysis.
What we certainly know so far is that at least for psychiatric disorders, these polygenic scores are not, or cannot be used to predict if you’re going to develop schizophrenia or not, but they could have some use, along with other risk factors, when clinicians are trying to determine the best course of treatment for the patient – so they can have some uses. But at present, the only useful polygenic risk scores that we can generate are largely based on individuals of Caucasian ancestry. And what the study of the East Asian samples told us was that if we want to generate polygenic risk scores for individuals of different ancestries, we need large GWAS based on those ancestors.
This is important because if you wanted to include polygenic scores as part of a clinical workup in a hospital in a major cosmopolitan city that’s made up of individuals with lots of different ancestries, it’s completely unethical to make that treatment, or those options only available to people of one ancestry, not another ancestry. So if we want to incorporate polygenic risk scores into the clinic, we need greater diversity of underlying GWAS. That’s why there’s a big drive to try and recruit and collect patient and controlled samples from different ancestries around the world. And not just to collect data and move it to European or US databases for analysis. There is very much a push towards trying to get local colleagues working on taking genetic analysis, and working in their own communities in their own universities around the world on this research.
FLG: Yeah, it’s an incredibly important area. So I think that kind of leads quite nicely on to my final question, which is what does the future neuropsychiatric landscape look like to you?
Derek: I think I remember reading a very good review article 10 years ago, and it was, I think, entitled, something like GWAS: The beginning of the end. And it was the idea that now that we could undertake GWAS, it was the end of the beginning. So it was making the point that in terms of the timeline through from gene discovery, through functional biology to translation for patients, we were just really at the end of the beginning, which was the gene discovery phase, and 10 years on, I still think we’re there, because GWAS hasn’t gone away, it’s actually bigger, it’s got more powerful. And what we’ve discovered is that complexity, in terms of the pathogenicity, and the many genes contributing to these disorders, is far in excess of what we could ever have expected.
So I still think that we’re getting towards the end of the beginning. And from here, then the major challenges really are about turning genetic findings into disease biology. And whereas again, we can find hundreds of risk genes, or maybe even 1000s of risk genes, there’s a lot of functional follow up work that needs to be done on those genes almost individually, to try and figure out how genetic variation affects their function, how that affects cell biology, and what the consequences then are for the brain and for patients that carry those variants in their genome. That then, again, is a basis for new targets for drug discovery.
I’m hoping maybe one day there might be some wins in terms of drug repurposing. So one thing that’s possible is when we when we know which risk genes are contributing to a disorder, there’s lots of drugs that have already been developed, you know, for use in other illnesses, etc. And it’s possible that some of those drugs may end up targeting some of the genes and proteins that we’re now finding to be involved in schizophrenia, that gives the opportunity to maybe repurpose those drugs for use in psychiatric disorders. Whereas, you know, previously that wasn’t thought to be sensible, or even a possibility. And I think that drug repurposing might be a way towards identifying some initial new treatments for patients, and kind of keep the momentum up to show that these genetic studies have the potential to improve patient outcomes. Because again, repurposing drugs is a quicker option to develop new drugs, but new drug development is something that needs to happen as well, but it’s just going to take a lot more time.
FLG: Yeah, I think obviously that would be fantastic. That’s actually all we’ve got time for today. But I’ve found that incredibly interesting, and I’ve learnt so much about risk genes and psychiatric disorders. I’d just like to say thank you so much for taking the time to talk with us today.
Derek: Thanks very much. It was my pleasure.
