Patrick Short 0:02 Hi, everyone, and welcome to the genetics podcasts. I'm really excited to be here today with Dr. Nicky Wiffen, who is a group leader at the Welcome Centre for Human Genetics in Oxford. Nicky leads the computational rare disease genomics group, which uses computational approaches to interpret the role of genetic variants in rare diseases, Niki, it's so great to have you today. Thanks so much for joining me. It's great to be here. Thank you, I'd love maybe if we could just start for people who aren't familiar with your work to have a little bit of an overview of what you and your research group are working on. And you know, how you got to work on rare disease genomics in the first place? Nicky Whiffin 0:35 Yeah, definitely. So our team looks at trying to find diagnoses for patients with rare disease. So traditionally, our approach is to well, when a patient ends up in a clinic, they have genetic testing done. And traditionally approach is to only look at a very small proportion of the genome, around one and a half percent, that directly encodes the proteins that go and do work around our body. With that approach, were relatively successful. In some diseases, we can diagnose genetically kind of 90% of patients, but in others, this is severely limited. So it's more like 20 to 30%. So we're clearly missing something. And we try and figure out a bit more about what we're missing. So we've focused on variants that are outside of this protein coding sequence. But our main interests are what we kind of term near coding, which are regulatory regions that are directly adjacent to these protein coding sequences that have incredibly important roles in regulating the amount of proteins that are produced. Our favourite ones are untranslated regions, UTRs, that are directly up and downstream of the protein coding sequence. And they do all sorts of roles like regulating the stability of RNA. So that's the kind of precursor before proteins are between DNA and proteins. And that RNA kind of persists in the cell and gets translated into protein. But these untranslated regions, regulate how long the RNA persists, and the rate at which it is translated into proteins, they have incredibly important regulatory roles. And we're interested in finding the subset of variants within those regions that actually can have severely detrimental effects and lead to disease. So what Patrick Short 2:05 in terms of kind of scale or real estate you mentioned, about one and a half percent of the genome is coding? What percentage then would be in these near coding regions that you all are looking at? Roughly? Nicky Whiffin 2:16 Yeah, so the thing that people don't realise is that UTIs are roughly equivalent to the size of the protein coding sequence for each gene, the majority of that is the three prime ends. So they're kind of downstream gene and which, and they're kind of smaller, the five prime ends are upstream as much smaller so that they can be a smallest side, I think the average is around 200 base pairs, but they can be much longer. And actually, we're finding a lot in our research at the moment that genes that are important in their kind of in that dosage. So loss of function intolerant genes tend to have much longer, more complicated UTRs that have more regulation in them to tightly regulate the translation of those important proteins. Patrick Short 2:51 So how much people listening may be familiar with the triplet code, we've got a pretty good although not perfect understanding of what a change to a protein coding section of DNA does to the ultimate protein, how much do we know about the code in these near near coding regions? And where are we at a few years ago? Where are we at today? And where do we need to be? Nicky Whiffin 3:13 We know very little don't know very little, I don't think we've made a huge amount of progress in that code, we need to do a lot of research on that. One of the things that makes it particularly tricky is when you look at a regulatory element, it acts as a whole. So if you're looking at the triplet code, and predicting the effects of a variant on the protein, you can put it that a single variant will change a single amino acid, whereas in, say, a UTR, a single variant might shift a structure which affects the regulation as a whole, or it might have absolutely no impact, but look like variants that have a big impact if they're in different portion of the ETR. So it makes us incredibly difficult. And you have to look at things like a UTR by UTR basis, Patrick Short 3:53 And I guess every gene Scott different size UTR is probably different structures within it. Are there are there supersets or groups where you can start to say, Are there any patterns that you can start to build up? Or is it really starting from a blank slate almost where we know these things have a function, but we're trying to understand the language so to speak now? Nicky Whiffin 4:13 Yes, there. ETs are incredibly variable across different gene sets, not just in terms of length, but also the makeup of different regulatory elements within them. Some of the variants we've most recently studies are ones that creates upstream start codons. So they create a just an ATG within the UTR. And that kind of causes the ribosome to pause and decide whether to translate early and can kind of disrupt the translation, the downstream proteins, these events that we know can be very impactful. But again, it's very dependent on where they occur and in what context they occur. So it might be that they occur but they occur downstream of the Start that's already being used, which means that they don't actually get seen by a scanning ribosomes that does the translating, or it might be that they are created into a context that actually isn't favourable to the ribosomes so it completely ignores it. keep scanning and actually makes it to the branching coding sequence. So there's a lot more complexity in just creating an ATG, it's what is that context? Is that created into? Patrick Short 5:08 What what are you starting to learn? I think one of the interesting approaches that you and your group take is you look at cases where something has gone wrong with with one of these near coding regions. So a child or an adult has a rare disease that is likely or potentially the result of one of these changes. And what you can do from this, I think, is start to understand by what happens when we break this element, we can learn a little bit more about the function of the element itself, maybe you could talk a little bit about how that process works, you know what you've learned so far, because sometimes there's a huge amount of value in in a small number of cases where you can really have a breakthrough almost where you understand by understanding what happens when it goes wrong, we can learn something fundamental about how these mysterious parts of the genome work as a whole. Nicky Whiffin 5:53 Yes, one of the things I find most fun about working in this area is it's not just about finding the diagnosis. It's also about finding something new about fundamental biology and kind of teaching. So the upstream start codons that exist and fight from at odds are a natural occurrence actually. So they create what we call upstream open reading frames or you off, and these naturally occur or that predicted to occur and around half of what known protein coding genes. And just give us an example of how little we know about these elements as I had a talk at the American Society of Human Genetics conference a few years ago. And it was great fun to stand in front of a crowd of metastasis and tell them that these elements exist in their genes they've been focusing on for decades that they've never even heard of before. But they're regulatory elements that are very, very common. But and we can kind of by looking at the variation and the way that leads to disease, we can look at when these elements are more competitive, the word, not the element itself being more deleterious where they where they have more of an impact on protein translation. So they're kind of sis arbitrary elements that down regulate protein translation by disrupting this ribosomes scanning. But if you find a variant in a patient that's deleterious, you can see okay, that's obviously a very high impact variant that's having a really big effect on downstream protein translation. But then you can look and kind of population controls and say, What are these genes? Are these particular types of variants don't seem to have such an impact on regulation. And from that those genetic variants you can actually learn when these elements are more regulative. And when they are kind of not having Patrick Short 7:18 such a big impact? And what's the estimated impact on patients and diagnoses? I know, you worked recently with Jamie Wallingford and a number of other scientists to look at clinical interpretation of variants in non coding regions. And I as you know, I've been also interested in this and what I did a lot of my research in, and it's still a an area of clinical practice, that's really very challenging. Where are we at today? How many patients do you estimate may have diagnoses that are not being made in most labs around the world? Because we're not looking at these parts of the genome? Nicky Whiffin 7:51 People ask me this question all the time. And I said, No, but I really, really want to have like to pull out my back pocket. Sadly, we don't have that yet. One of the problems is kind of, it's hard to do these analysis to get a real number on that, because there's also so much noise in the non coding space. I mean, a lot of work you did during your PhD was had the conclusion that it's individual basis that have an impact, it's not entirely entire regulatory elements. So you've got to find ways of filtering the variants that you only got the ones that have a large impact. And if you don't do that properly, you end up with a lot of noise. And one of the ways in which you were looking at this was to look at De Novo variants, so ones that occur only in a child or that hasn't affected parents, so they have a higher prior of being a disease causing, and we have some work we're doing in the genomics England 100,000 Genomes Project at the moment where a superstar postdoc in the team Alex Geary is working to try and look at both de novo and inherited variants using we think, hopefully, some clever approaches to try and finally put a number on that what proportion of patients have a likely diagnostic UTR variant or promoter variant, or other kind of regulatory variant, but at the moment, we don't really know what we're expecting to find. And that can the most progress in this field has come from single gene disorders. So patients where the clinician can say that a patient has a variant in that gene. But in a lot of the kind of developmental disorder space, for example, there are so many genes that could be the cause it's much more heterogeneous, which means that you can't really pinpoint exactly which gene you're expected to find a variant in. But there's some kind of nice poster child genes. So we did some work recently with Caroline right, and the develop department developmental disorders team Messiah. And we kind of found this gene called map to see where actually we find that 25% of patients in DDD with a map to see variant have won in the non coding region. So for some genes, this non coding variants are incredibly important, but for others you can diagnose so 90 plus percent of patients in the coding sequence so it's very, very gene region. Everything depends on it. So it's very hard to put a number on that. Patrick Short 9:57 Do you have a sense of whether that gene enough to see is more of an exception or more of a rule. Do you think that if B and I, I have asked you to speculate here, I'd be great if you had the data. But yeah, curious whether that is, is, is an outlier or not Nicky Whiffin 10:14 as high as 25%. I think it's definitely more of an exception than a rule. I think most genes, it's going to be lower than that. But it's, it's not going to be the only one. And it's some kind of combination. I think, I think the main factor for this is probably because it's very doses sensitive. So even severity, that in some genes we wouldn't think would ever be that deleterious in math, you see, it seems that they can cause a bit of severe developmental phenotypes. So that also factors in it. And also just the fact that we had coverage in the sequencing data to actually be able to profile variants in that region, which is obviously very important to what we can detect because DDD, as you know, is based on exome sequencing data, so only capturing the protein coding regions. However, I was talking to some people yesterday, however, that's a bit of a misnomer, because whole exome sequencing does not sequence the whole exome. Lots of exons that are non coding, though whole exome sequencing is a bit of a misleading name for the technology. Patrick Short 11:07 When you and I were both earlier in our careers, there was an honestly term still used, but this term of junk DNA of the non coding genome being 98%, or 80%, depending who you ask junk. I think we're learning that that's definitely not the case. But I'm curious what piqued your interest in the first place of studying something that a lot of the field said was, you know, basically, the actions going on in the protein coding genes? Sure. There's, there's some important stuff. But what what made you interested in focusing on the part that nobody else was, was actually all that interested in? Nicky Whiffin 11:39 Yeah, junk DNA term just kind of highlights how we'd rather discount something that we don't understand. Yeah, then realise, we have to put a lot of effort into trying to understand it, my PhD was actually on doing a lot of genome wide association studies in colorectal cancer. And obviously, a lot of GWAS hits are found in the non coding regions, I found that kind of quite unsatisfying, because I wanted my research to have more of an impact on patients, which is why I kind of shifted away from GWAS and into the kind of rare variant rare disease fields. But I think it's part of that that's combined my interest and also just kind of frustration a bit in the fact that we ignore. So we've spent a lot of effort and trying to find the genes that are involved in certain diseases, but then we ignore half of the gene. So we actually only focused on the bits that code for protein, but we actually know that these genes are important in this disease. So we should be looking at the regulatory elements for those genes as well. But the fact that we don't is, is mainly because we just don't know how to if I can do a little bit towards trying to teach us how or make some difference in that way, then I think that's a very rewarding thing to do. So that's kind of how I got to this space. But Patrick Short 12:45 yeah, and I think you all are doing, you're doing really important work. And I think genomics England is doing important work here, where they're sitting in between clinical practice and research, because I think one of the challenges has always been if you're running a clinical lab, and the question is do I do I charge somebody $1,000, to do a whole genome, or 300, to do a whole exome. And we and we only know about the whole exome. So that's all we can interpret, then it's really clear, there's not a justification to insurers, the National Health Service, or whatever to say, we're going to sequence all this stuff that we really don't know anything about, and we can't report on, but you end up in this catch 22, where if you're, if you're not sequencing it, then you don't know what we're missing. And so most of the clinical sequencing around the world is either single gene and multi gene panels are starting to be exomes. But what I think genomics England and a few others have done to really start to break this cycle is say, we're going to do whole genomes, because we don't know what we don't know. And by forcing the issue, we can actually really figure out, you know, is this going to, is this going to be worthwhile in the long run. And I think starting to chip away at that and say, How many diagnoses can we get out of this in depth sequencing? And I think when you do the math, especially as the cost of sequencing comes down, it's going to start to make sense very soon. If it doesn't already, that we should just simply be whole genome sequencing and everybody, right, even if you can boost diagnosis by five or 10%, the impact of that is so enormous on the patients and the costs that I imagined the numbers start to add up. Is that your view on it as well? Nicky Whiffin 14:14 Yeah, I kind of think there might still be some diseases where panel testing makes sense. But it may well be expanded panel testing. And if we know if we have a set of standardised regulatory regions that also should be included within that panel, that it might make sense. And if we can diagnose 99% of patients using an extended panel, then there is no reason to go down or genome sequencing one time where it really can make an impact as where we have these long diagnostic odysseys for these patients. And if you're going in and you're doing karyotype, in then you're doing panel testing, when you're doing exome, then you'll finally get into genome that's a long time for all of those processes to work out. And that's just adding to the kind of uncertainty for these patients. Whereas if we go into the whole genome, we can still focus our search first on those on kind of expand it as we go along with the kind of more likely candidates first but the data is already there. And it In the long term, you're saving money. And you're also kind of saving a lot of kind of time, both for the patient and also for the kind of scientists doing the work. Patrick Short 15:08 Have you all looked at other kinds of sequencing technologies, long read sequencing or anything like that? And is there value to something, something beyond the current short read sequencing that we're all more or less used to working with? Nicky Whiffin 15:22 I think there's definite folly, but I have not looked at any long wave data. And one of the kind of really interesting things is being able to do phasing. So it's not necessarily having to have access to parents to be able to know which parents something was diagnosed, it's to be able to find compound heterozygous variants were pretty Patrick Short 15:39 good explain. Not everybody. For those who don't understand phasing, I think it'd be a really important concept. Maybe you could jump in and explain what that is. That was a big bit of jargon. But I just don't know, it's great. It's really good. I don't think we've had an episode where we've really talked too much about it. But it is, I think it's such important point, especially in rare disease. Nicky Whiffin 15:56 Yeah. So this is where you're kind of you inherit one copy of each chromosome from each parent. And it's difficult when you're only sequencing very short sections of that, to figure out whether a variant that is 500 base pairs away from a different variant came from the same parent or came from a different parent. So what phasing is trying to do is trying to predict, and computationally which variants arose on the same like what we call haplotype. But from the same parent, and which ones up came from the different parent, this is important when we're looking at a recessive disease, where we need say, both copies of a gene to be knocked out. And if you see two variants that you'd expect to knock out a gene on the same haplotype, then that's only gonna knock out that single copy. Whereas if they occur on the different, different haplotypes, they've been inherited from different parents, then you get both copies knocked out. So one of the ways we can find this out, it's by sequencing the parent, and you can see which parent had which variant. But that's obviously as the cost you're doing three times the amount of sequencing, whereas if you can do phasing, you can find all along with sequencing, you find this out just from the patient themselves, Patrick Short 17:01 having access to face data and understanding not just the child's genome, but both parents, I imagine is really important for what you do, right? What do you have a sense of how much that improves the diagnostic rate, or how important it is to actually have that complete family information versus just just data from an individual? Nicky Whiffin 17:19 Yeah, I can't remember the numbers that go with this. But there's been lots of studies that have shown how the diagnostic rate is far higher, if you have access to a full trio, or at least some family structure, rather than just having a sequence of the patient. And one of the reasons for that is we can identify these de novo variants, which we only find a small number in each patient, but they have much higher prior of being disease causing when you have unaffected parents and an affected child. The other explanation for that is recessive disease, like we just spoke about. But that's much higher when you have say, related parents, then you can kind of hypothesise that will be a recessive cause. But again, as we just spoke about, you get that phasing information by doing at the tree. So trees are incredibly valuable. But there are other family structures available as well. So you might have two affected children, and then you can have a quad. And even even having anything other than just the single patient can be very Patrick Short 18:12 valuable. You've done quite a bit of creative work so far in your career, I was always, you know, very impressed with the open reading frame work, you were doing exactly like you said, this is something that nobody really actually was on anybody's radar. But there's actually something really interesting there. There's another one of your papers that actually bring up a lot because I think it's another kind of non intuitive but really creative example of genomics and drug discovery, which is some of the work you did as part of the Nomad Consortium on loss of functions. And lock two would be really great to hear from you about how that came about. And and some of the lessons you learned from doing that work. So I think it's it's very, it's initially non intuitive why it's so useful. But then I think once you dig into it, it certainly for me was the kind of aha moment about how genomic datasets can be used in drug discovery in really creative ways. Nicky Whiffin 18:56 Yes, this project was the kind of typical science example of some of the big papers come from being in the right place at the right time. This was a project that Danny McCarthy had already initiated with 23andme, where they were interested in kind of sharing the value of humans as an experiment for drug discovery. So Parkinson's of luck to you is quite an interesting example here. So gain of function mutations in Mark Two are known to cause Parkinson's disease. So because of this huge increase in interest in the pharmaceutical world, of trying to develop inhibitors to lock to as a treatment for Parkinson's, but when people have kind of modelled this in mice and other model organisms, they've seen some very severe lung, liver and kidney phenotypes in the kind of knockout animals. As we know, these animals are our models, but they're not they're not actually humans. But there's nature does this wonderful experiment where it introduces all this natural variation, and what can we use for that natural variation, to look at what the effects of knocking down a specific drug target might be in humans, and there's a great companion book to this as well by Eric medical that kind of looks This on a more broad scale and luck to you as a specific example we were looking at. So what this allows us to do is to look at these loss of function mutations. So where we kind of predicted to lose half a copy of, well, half the amount of protein, that's what the downstream impact of expect and look in kind of large scale biobanks or datasets with linked health information and say, Well, what impact does that have in these people. So this was kind of the paper full of completely negative results. But that was exactly what we wanted. So I've got all of these flatline figures that work, they're quite fun to create, but showing that there aren't, there doesn't seem to be any increase in any lung, liver, kidney, or other kinds of severe phenotypes, as far as we could test with the power that we had in humans that have half the amount of luck to your protein. So what this suggests is knocking down lots, you by half shouldn't be deleterious in when it's done in clinical trials. So this was really kind of positive, positive negative data, which is a really great thing to show. Patrick Short 20:57 Yeah. And I think, as a general kind of class of research question, I get some advice very early on in my PhD, which was, if you focus on a question or set of questions that no matter what the answer is, it's going to be interesting, then you'll be fine. Where you run into trouble as if you have a question that is only interesting, if this one answer or the other. And then as a scientist, you're, you're stuck with either getting the interesting answer, or they're getting the non issue. And this is a perfect example of that, where if you'd found that people with a partial knockout have flocked to end up with significant kidney, liver lung disease, then that's a really useful piece of information, because you can tell the world, hey, we actually do need to be careful inhibiting this protein. And if you find the reverse that actually we don't see much evidence that there's some people who have this knocked out their entire life and every cell of their body, like you said, by nature, and we don't see kidney disease showing up in their medical records or things like that, then it doesn't mean it's definitely safe. But it's a pretty good indicator. So it's a really good example of that principle work. Nicky Whiffin 21:58 Really important caveats. And one of those being that if you have it knocked down from birth, there are potentially like response mechanisms in the body where they can kind of correct for it by maybe up regulating something else, which is not going to be happen, possibly, if you give somebody a drug later in life, but as you say, it is kind of lifelong knockdown, whereas the therapeutic is only going to be later in life. So it there's kind of some it's not the same experiment, but it is this kind of promising, promising data. Patrick Short 22:27 Yeah, that is a really important and interesting caveat, essentially, that there may be some kind of compensation where if you always have half the protein, then the body figures out a way to make it work. But like you say, if you introduce it instantly, at some point later in life, you you can't rely on that compensation unnecessarily happen. Yeah, Nicky Whiffin 22:44 exactly. And we also obviously can't predict off target effects that the drug might have very easily. Patrick Short 22:49 And you recently on the topic of advice to early career researchers, you recently started a research group. I'm really curious how did it feel day one walk into the new office? I'm sure it was just you on day one. But you had a plan to build the team. How did that feel? What What brought you to that moment? And how's it been? Nicky Whiffin 23:05 I guess one of the most disorientating things is I didn't walk into any office, I stayed in exactly the same office in my own house. i Yeah, it started in September 2020. So very much having delayed from June 2020, because I didn't feel like starting a lab in the middle of a pandemic, which shows you a bit of our short sightedness of how long we thought this thing would go on for. But yes, I started in September, and I think gotten how long it was, I think it was nearly a year before I actually went into the office and found out how tall my team members were. So given I managed to delay for that amount of time, I've actually set up some recruitment. So actually, I only had a couple of weeks of being the only member of my team, and then very quickly had a D feel student and a postdoc, and actually an intern during the group. So that was that was quite nice. Science can be quite an isolating place anyway, you don't want to sit in your own house as the only member of your team for too Patrick Short 23:55 long. How has that transition been for you personally, what changed from being a postdoc where you were obviously working as part of a team and had often postdocs to work really closely with PhD students and others, but from making that transition then into really being able to set your own research agenda in a way you probably weren't able to before? What was that like? And what were you expecting? And how was it? Nicky Whiffin 24:17 Yeah, it's great fun, actually. And I yeah, I learned to code during my PhD, I'm not gonna say that I'm the best coder in the world. So it's great to be able to employ people who are better at coding or better at coding than me to actually do my ideas. But one I love kind of being able to create the environment that I want to work in and making things kind of support supportive and inclusive, and I really enjoy that side of things. The kind of key difference I think in the way that you work is as a student or a postdoc, you get to kind of say, oh, today I'm going to work on this and you get to kind of shut everything down, else down for the day and just work on a soul bit of code soul question for the entire day. As a pie. It's very much a jumping up out every half an hour trying to reprogram your brain to kind of go from one task to about 20 different things that are on completely different topics is quite a challenge. I found that incredibly disorientating to start off with but I think I've kind of got that one a little bit sorted by this point. Patrick Short 25:15 And what are you most excited about right now? What are the areas of research you and your team are working on. And also, I'd love to hear about any areas that maybe aren't what you're working on, but as a whole you think are exciting for the next coming years that especially ones that maybe aren't on people's radar that you think in a few years, we'll be saying maybe like the open reading frames a couple of years ago when nobody was talking about it, but then it starts to it starts to become a lot more of a topic, what what are you most interested in, in, in the group and in the industry as a whole. Nicky Whiffin 25:43 We kind of have quite a broad focus in the group from the very, very clinical stuff. So we've just put out a set of guidelines for how clinically we should interpret that to the non coding regions. So that was a really great collaboration with a lot of awesome people that think, think in the space. But also, as I said, we're trying to put a number on this, how many patients can we expect to diagnose through using the genomics England data, but also we're trying to look more on the regulatory side of things and really understanding UTRs hoping that that will help us to interpret variants down the line. So there's a wonderful detail student at Homer who is looking at kind of how UTIs vary across different genes. So as we vary the tolerance, the loss of function, we see massive differences in the makeup, the length and the makeup of UTIs. And also, if we look at as a developmental disorder, dominant development disorder genes, they look completely different to to kind of genes on a whole. So there's important things to learn about regulation, which hopefully downstream will help us interpret Barian. And we've kind of got this big a collaboration with them. In order to secure a Danish company, and Joe housing, who's the head of genetics to look a bit more about, you often need to verify their role in more common phenotype using loved UK Biobank data. So really excited about that kind of on a whole set? It's an interesting question, I think we're going to learn a lot from having the whole genomes in all of the UK Biobank cohort, I'm really excited to see what people do with that data, I think there's gonna be some really awesome things we can can learn, but also some really cool computational and analytical approaches that people come up with, with how to kind of wrangle that data and learn new insights out of that. So I'm very excited to see where that takes us. Patrick Short 27:20 I completely agree. Do you think that on the topic of both whole genomes and large scale data sets and also understanding more about the non coding regions? Do you think there are enough people on the planet for us to to actually understand these by looking at human data? And what I mean by that is, you know, if we were to sequence whole genome, everyone on Earth, it's I think it's possible that the complexity in these regions is still too high that we couldn't using data from 6 billion people alone, really pick out which bases are important, which ones are not important, which ones cause disease. I know a lot of people agree with that statement. But a lot of people vehemently disagree with that statement. I'm curious where you fall on that, and and then maybe we could talk about some of the other strategies to tackle that problem. Nicky Whiffin 28:02 I think one of the important things is that you don't have to observe a variant to tell something about it. When you get to a certain number of people that we've sequence. And when you look at variants that are in positions that are highly mutable, the fact that you don't see that variant is as informative as if you see it. So that implies that it's not compatible with life. So with variation that we already know, or is that saturation at the sample sizes we have now. So I'm talking about C to T changes at CG dinucleotides, that are that are methylated, we know they have a very high mutation rate. And just by looking at those votes that you don't see, you can learn a hell of a lot. So I think the kind of the fact that humans have been under selection for how many number of years means you're not purely looking at a DNA sequence in isolation, you're looking at kind of what's happened in the past. And by combining if we had 6 million people, the data that would be amazing, I think we would be able to learn a hell of a lot from that. Patrick Short 28:57 Yeah, I completely agree. And I guess it's, I need to update my numbers. It's probably 7 billion now. Right? 6 billion is what it was when I was a kid. But that's the number that's burned into my mind, but no longer. Well, thank you. I think it's great note to end on. I really appreciate you taking the time and talking about your work and being with us today. Nicky Whiffin 29:14 Well, thank you very much. It's great to be here. Patrick Short 29:16 So if you want to follow Nicky and her work, she's on Twitter, at Nicky, within you got the first name last name, which is great. That means that you were you are a little bit early to the party and I know Nicky Whiffin 29:26 A very obscure name. Patrick Short 29:28 Yes, that's right. I couldn't get the Patrick short, unfortunately too, too common. And I think you're hiring right now. Right? So if people want to learn more about your your group what you do at your NICU if and on Twitter, but I think your website as well as rare disease genomics.org. Maybe you could talk a little bit about who you're looking for and what they'd be working on. Nicky Whiffin 29:46 Yeah, definitely. The best thing to do is to find the contact details for the through the website, or just by Googling me again, obscure name means I'm very good, global, slightly terrifying if I ever do anything well, but we're looking for kind of a senior postdoc to lead collaborative just never noticed that I briefly mentioned looking at UTI variants and using kind of multi omics within the UK Biobank. I think it should be a really awesome project. Patrick Short 30:07 Awesome. Well, great, thanks. Thanks sticky and thanks, everyone for listening. As always, we'd really appreciate if you could share this episode with a friend if you liked it and leave a review on your favourite podcast player. This helps other people find us. So thanks very much for your time and we'll see you next time.