Welcome back. This is the last regular module in this unit on schizophrenia. And in, in this module we're going to talk about an alternative to looking at SNPs, what are really called rare variants. And just to recap our efforts, to identify the specific genetic factors underlying an individuals risk for developing schizophrenia, the initial candidate stu, gene studies did not work out. Recall the Sanders' studies that we talked about. GWAS has, there's been real significant progress with GWAS. We can identify variance. Recall that the last major GWAS identified over a hundred variants. But their effects are really small. That is, they don't in, if you inherit one, it doesn't increase your risk for schizophrenia all that much. So they're hard to identify. You need these really large samples to identify them. And most of the heritability remains missing. We don't know where it is. GWAS is designed to detect the importance of a certain type of genetic variant. SNPs, those single nucleotide polymorphisms, that actually are common in the population. That, they have a frequency of less, at least 1% in the population. So GWAS misses certain types of alternative genetic variants. First of all, rare variants. Variants that don't occur very frequently in the population. Variants that occur maybe one out of 1,000 or one out of 10,000 individuals. They're really not assessed in GWAS. And in part they're not assessed because they're kind of hard to find because they, they're rare. And the other thing that's really missing here, or another thing, there's more than this, are structural variants, the Copy Number Variants that we talked about earlier in unit four. Now it turns out that it's, it's actually pretty hard to study rare variants because they're rare. But at, but in this module what I'm going to try to do is just convince you that rare variants are also probably pretty important in understanding the genetics of schizophrenia. And I'm going to do it really by illustrating two phenomena. One involves Copy Number Variants. The first involves what I think is really a fascinating phenomena, which is the paternal age effect. Something I, I don't think I really knew much about until about ten or 15 years ago. It turns out that many disorders has a paternal age effect. And that's true for schizophrenia as well. Here's the frequency of schizophrenia in a very large population. This is actually coming out of Denmark where they've surveyed in the, in the registry system there over three million individuals. So they have a real large sample. And they're just plotting here the rate at which the individuals in the population have schizophrenia as a function of the age of their father, which is in red, or the age of their mother, which is in blue, at the time of conception. And it's normalized so that the, the rate here is one if the father and mother are between 25 and 29. So they just normalized it that way. Now it turns out that there's a little elevation of risk of schizophrenia if you have a very young mother. I'm, I'm not going to talk about that here. That's probably reflects some other, maybe even nonbiological process, or at least nongenetic process. What I'm mostly interested in is this age effect here that occurs in dads but not moms. As dads get older, your risk for schizophrenia increases. It's not overwhelming. It isn't like 25% of the children of a 50 year old dad are going to develop schizophrenia. It's more like maybe 1.5 or 2%. But right, you can see there's a marked age effect here, depending upon the father's age on your risk for schizophrenia, but not your mother's age. Why is that? Well, it turns out there's quite a few human genetic disorders that show this pattern of increasing risk of the disorder with father's age, but not mother's age. Disorders like achondroplasia, a form of dwarfism, neurofibromatosis, progeria. Progeria, this is a, a, a, I think, a toddler with progeria. Progeria just means accelerated aging. And so these are individuals who look wisened by the time they're three years old, and they're probably, unfortunately and sadly, dead by the time they're ten years old. The it, the chance that an individual has a child with this devastating illness increases with the father's age, but not the mother's age. Behavioral disorders, we already saw schizophrenia, bipolar disorder increases with father's age, autism and intellectual disability. These are all disorders that have a genetic component to it. Could there be some genetic explanation for this? Well there does appear to be genetic factors that can contribute to an increase frequency of disorders as a function of father's, but not mother's age. We already know that there's a reason why genetic in certain genetic disorders would increase with mother's age. Those are disorders due to aneuploidy, something we talked about earlier in the course. Things like Down's Syndrome and extra chromosome 21. That increases with mother's age not father's age. These disorders are different. Why might they increase for a genetic reason? Well, it turns out, and I'll explain this graph here in a second, it turns out that in each generation we inherit about on average, and it varies a little bit from individual to individual, but about on average 60 new mutations. These are mutations that our parents didn't have but we do have. So they are mutations that arise in meiosis. And because they're new, they didn't, they don't exist in our parents, but they exist in us, they're called de novo mutations. Most of these mutations don't have any phenotypic effect, but every once in awhile they will have an effect. They'll be in a gene that's important and maybe you end up with progeria. It's not going to happen very often, fortunately, but it will happen. It turns out that most of those 60 new mutations that we inherit in our DNA sequence come from our fathers, and not our mothers. And the reason for that is, sperm are constantly turning over. They're constantly going through meiosis. And each meiotic division is an opportunity for an error in the DNA replication process. So each time there's a meiosis there can be a copying error. Mothers are actually born with all their eggs. They're not constantly going through meiosis like sperm is. So that type of error isn't going to occur in the mother's eggs, but it will occur in the father's eggs. This is a paper published in two, these are results from a paper published in 2012 in nature. A researcher, a group of geneticists, did a very careful analysis of this where they actually looked and determined how many mutations were being donated by the father versus the mother as a function of the father and the mother's age. In fathers, the number of the new mutations increase two per father's for each year of the father. So another way of thinking about it, a 50-year-old father versus a 25-year-old father, that's 25 years difference. So the 50-year-old father is going to be passing on 50 more mutations to his daughter or son than a 25-year-old father, on average. Now again, most of those 50 mutations probably don't have any effect, but every once in a while they will have an effect. Mothers don't actually, the, the number of mutations they don't, donate do not increase with age. So the increased risk of schizophrenia or autism or bipolar or progeria or hemophilia with paternal age is thought, at least in part, to be due to the fact that as fathers get older, they're more likely to transmit rare de novo genetic mutations to their offspring. The second phenomenon I want to talk about here. Are structural variants. Specifically copy number variants. And these have also been associated, they're rare, they tend to be rare. They associate with schizophrenia risk. [SOUND] The most famous of these is what's called Velo-Cardiofacial Syndrome. Velo-Cardiofacial Syndrome is variously defined, there are various labels for this I put them in the, in the note here to this slide. I think the proper term is 22Q11.2 deletion syndrome, I'll call it Velo-Cardiofacial Syndrome, It's also called the George Syndrome and Spritzen Syndrome. What Velo-Cardiofacial syndrome is, a developmental disorder that causes cleft palate, that's the velo, as well as cardiovascular abnormalities, thus the term, Velo-Cardiofacial syndrome. Individuals with this syndrome also often have learning disabilities. As copy number variants go, and remember copy number variants are regions of the DNA that are at least 1,000 bases long but could be multiple millions of bases long, where we inherit other than the normal two copies. We might heri, inherit an extra copy, so we have three copies or we might be deleted so we only have one copy. In the case of Velo-Cardiofacial syndrome, it occurs in one out of every 3000 births. So it's very frequent as far as copy number variance go, and it's it's due to a deletion of a bout 3 million basis, that's a lot of basis of DNA, 3 million basis on chromosome 22. It's about 30 genes map to that region. So what's going on here? These are de novo mutations. That is, if you look at the fathers, they have the whole chromosome 22. And if you look at the mothers, they have the whole chro, chromosome 22. But in one of those, in this case, I've illustrated for the father, during meiosis, there was a deletion of 3 million bases of DNA on chromosome 22. So the child is missing those 3 million bases, they have the other copy but they only have one copy of all those genes. It turns out in this case the, the deletion is equally likely to come in this case from the mother or the father. Velo-Cardiofacial Syndrome is associated with schizophrenia. 20-30% of the individuals who inherit this deletion on chromosome 22 will develop schizophrenia. 1% of the individuals in the population with schizophrenia have this deletion, have Velo-Cardiofacial Syndrome. So it's a major factor in developing schizophrenia. I said there about 30 genes in the region that are del, deleted. Here's a list of all the, all the genes, or at least most of the genes. You might think that, so again, what's happening here? People with Velo-Cardiofacial Syndrome have one copy of these genes. Most of us have two copies. So presumably having one copy, somehow can lead to schizophrenia. So this region has been intensively studied to try to understand if there are genes in the region that might be implicated in risk for schizophrenia. Two of them have been studied a lot; this PAH gene and the COMP gene. We still don't know why deleting these 30 genes makes it, increases your risk 20, 30 fold for developing schizophrenia. But a lot of the attention, at least at this point in time, has turned on these part, two particular genes. So copy number variants are rare, like Velo-Cardiofacial Syndrome. And we know they are also, at least some of them are also, associated with risk for schizophrenia. And they're not being essayed in the GY studies. So it might help explain some of their irritability. Velo-Cardiofacial Syndrome is an example of one type of copy number variance. There are actually others that have been identified over the last few years. Here's a, a table not, not listing all of these, but listing a few of these that have been identified. It turns out that Velo-Cardiofacial Syndrome, the, that deletion of 3 million bases on chromosome 22, is specifically associated with increased risk of schizophrenia, not increased risk for other psychiatric disorders. But as human geneticists have looked at copy number variants on other chromosomes, what they have found is that these disorders, or these copy number variants, either when they're duplicated or when they're deleted are non-specific risk factors for multiple psychiatric disorders. And in fact multiple neurodevelopmental disorders, I'm just going to highlight one here, it's, it's probably the most famous in this particular table, 16p11,2. I don't actually know how frequent it is, and I don't know that we know how frequent it is it, that is, how common it occurs in births. But we know that if you're deleted this region on chromosome 6, 16, you're at increased risk for autism and increased risk for developing intellectual disability. In fact, about a half a percent of individuals with autism are deleted this region of chromosome 16. The same region if it's duplicated, so here you have one copy. Here you have three copies. Being duplicated for this region is associated with increased risks for other neuro developmental disorders. It again is associated with increased risk of autism. So having one copy of this region and as well as having three copies of this region increases your risk for autism. And there's a lot written on this region in terms of autism. But it's also associated with increased risk for intellectual disability, epilepsy, and indeed, schizophrenia. One of the fascinating things that is emerging in recent human genetic research on structural variances in psychiatry are these comorbidities again. Or you might call it plyotopy. Deletions and duplications of regions of our genos can give rise to rather distinct neurodevelopmental disorders. Why? We don't know why. We don't know why they're giving rise to these disorders, nor do we understand why sometimes it ends in schizophrenia. Other times autism, other times epilepsy. That's really on the, on the forefront, the frontier of current human genetic research in this area. To summarize what we know today about our circuit 2014. About the genetics of schizophrenia. Human geneticists call this, sometimes, understanding the genetic architecture of a disorder and by genetic architecture they mean, what do we know about those genetic factors that lead to risk for the disorder in terms of how frequent they are. And that's being plotted here in this two dimensional space. And how large an effect that variant has on your risk for the disorder. So that's what's meant by genetic architecture. And what's plotted here in this review from 2014, are the known identified risk factors, genetic risk factors, for schizophrenia. There are two types plotted here in this plot. There are the snips that have been identified in GWAS, these are things that are relatively common, I told you that GWAS looks for things that are common, that probably all of us are carrying some of these just not enough to develop schizophrenia. They're common, they have a high frequency, but they have a very small effect the effect here on the scale is very, very small, therefore they take gigantic samples to identify them. The other type of variance that have been identified, are these rare variance that are uncommon, they have a larger effect or like Velo-Cardiofacial Syndrome or that 16P duplication. They have a much larger effect, so they're plotted up here. But because they are rare, they actually don't account for a lot of the variances in schizophrenia. So we know these, we know these, but together they still leave most of the heritability of schizophrenia missing. We know more about the heritability of schizophrenia today then, let's say, a year or two years ago. But still, in all honesty, much of it is missing. [SOUND] Where are those other variants in this plot? Well, first of all, we know where they're not, at this point. They're not here. What would be plotted here are things that are common that have a moderate or large effect on your risk for schizophrenia. We now know those don't exist. They couldn't possibly exist, because if we had a sample of almost 40,000 people with schizophrenia, we would've identified them at this point. And we don't find them. They don't exist. So we know they don't exist here. We also probably know that these don't exist up here, rare things that have a real large effect on schizophrenia risk. If we're looking for the missing heritability of schizophrenia, where those variants lie is below this curve, not above. If they lie above the curve, we would have found them. We haven't found them, they probably don't exist. They almost certainly don't exist. So what are the ones that are going to account for this missing heritability. There, there are either rare variants that don't have a large effect. They might have a moderate or small effect, or there are common variants that have a really small effect, much smaller than what we're seeing here. In either case, it's going to be very, very hard to identify these. It's going to take massive samples, lots of resources, lots of research time, lots of money for doing genotyping. People are debating whether or not it's worth it. To the extend it's worth it, and I, I happen to be again, as I said before, in the, of the, of the mind that it is extremely important to do this. The argument for doing it is that it's beginning to tell us something about the path of physiology of schizophrenia that ultimately we hope will lead to effective treatments and preventions. That's the argument for doing this. The argument for, for against doing it is you're getting these lar, such small effects, are we really going to learn much biologically? I think you're going to see this debate played out over the next five or six years. And maybe five or six years from now, I at least hope we would have an answer to this question. So, this comes to the end of the regular modules for Unit 5. We talked about skino, schizophrenia,. It's a relatively common psychiatric disorder, 1% of the adult population suffers from it. It's a devastating mental disorder. Twin and adoption studies have shown us that is a highly heritable disorder. The environment is still important, but the nature of environmental influence appears to be predominantly of the non-shared environmental, not the shared environmental kind. We went through that in this module. What's the important non-shared environmental factors? They appear to be primarily non-specific neuro-developmental insults. Insults to the developing brain. Things that happen in utero, perinatally, or very early in neuro-development. What do we know about the nature, the specific nature of the genetic influence? Again, it appears to be highly heritable. There are many genetic factors that appear to be contributing to our risk for schizophrenia, just like two individual differences in height. There are probably thousands of these in the genome. We've found maybe a little bit over 100 at this point. There are many, many more that could potentially be found if we want to invest the resources to do that. The ones we've found, if they're common, they're coming out of GWAS. They have a very small effect on the phenotype. If they're rare variants, they have a larger effect, but they still don't account for much of the variants because they're rare. In total, if we aggregate everything we know, most of the heritability remains missing. And that's true of, again, of every human genetic complex phenotype today. Thank you. [SOUND]