 All right, we're going to start. Our next speaker is David Page from the Whitehead Institute, who will be talking about sex chromosome evolution and medicine, David. Okay, thank you, Eric, and thank you very much for the invitation to speak at this terrific symposium on a wonderful occasion. And I would like to begin by thanking NIH and NHGRI for 21 years of support of our studies of a chromosome that was in need of some respect. Studies that have allowed us to completely rethink the Y chromosome structure, function evolution, and medical significance. And what I want to do today, so yeah, I should say that I've spent the better part of my career pondering the wonders of this pair of chromosomes. On the left, the stately and upright X chromosome, and to its right, somewhat downtrodden, even demure Y chromosome. I've spent the better part of my career defending the honor of the Y chromosome, thankfully with the support of NHGRI, in the face of innumerable insults to its character and its future prospects. So there's so much I'd like to tell you today, but I've decided to be compact in my messages and perhaps dive into a little bit more detail. What I'd like to say today, I'd like to make three points. First, I want to tell you that sequencing the human Y chromosome has yielded completely unanticipated insights, including, as I'll explain today, a possible mechanism of origin for the chromosomal defect in Turner syndrome, which, as you know, exclusively affects girls and women. Second, I will point out that due to the structural complexity of the Y chromosome, sequencing it required some unorthodox strategies. Strategies that would allow us to differentiate between repeated sequences or paralogs that are more similar than alleles. That is, we had to travel beyond what I might call the allelic limit. And third, I will recommend that a similar strategy be combined with the latest sequencing technologies to scrutinize about 160 of the most structurally complex eukromatic sites on the human X and a lot of cells. These are about 160 sites that likely account for a disproportionate burden of disease. And I will suggest that this investment will be returned each of the, I think we heard this morning, million times, or did Eric say a billion times that the genome is resequenced. Okay, so let me first give you a crash course then in how we think about sex chromosome evolution. What we've learned over the last 15 years is that the human X and Y chromosomes evolved from a perfectly ordinary and unsuspecting pair of autosomes. So as I'm sure you discussed when you get together on family occasions, you'll recall that 300 million years ago, when we were reptiles, we had no sex chromosomes. We had only ordinary autosomes. And what happened beginning about 200 or 300 million years ago was that one of a perfectly ordinary pair of autosomes sustained a mutation that would begin to give rise to what lives on today as SRY, the sex determining gene on the Y chromosome. And what transpired, and I should say this is work, this understanding of the sex chromosomes is based completely on genomic analyses over the last 15 years. What then transpired was that first in the immediate vicinity of SRY, and then over a larger region of what was becoming the Y chromosome, crossing over between the X and the Y was suppressed. And it was actually the suppression of XY crossing over that allowed the X and the Y to differentiate from a perfectly ordinary and identical pair of autosomes. Now, sexual recombination is a good thing for the health of genes in the long run. And so the X chromosome retained the rejuvenating benefits of sexual recombination due to crossing over in female meiosis during oogenesis. But the Y was now left to reproduce clonally without the benefit of sexual recombination. And so it began to lose genes. And in this expanding blue region, the Y chromosome lost genes. If you take this out, if you sort of extrapolate this out to infinity, it does not bode well for the Y chromosome. And thinking of our timing about 10 years ago, actually nine years ago, this was taken to its logical extreme in, of course, a journal that we seem to be making frequent reference to today, this obscure one called Nature. And in an editorial somewhat grandly entitled The Future of Sex, what I just told you was pointed out. That is, the Y chromosome is particularly vulnerable against the ravages of evolutionary time because it doesn't have a matching partner with which to swap genes so it cannot retrieve lost genetic information by recombination. And then came the really devastating conclusion in this editorial, the present rate of decay, the Y chromosome will self-destruct 10 million years. Well, a graduate student came running into my lab having read this, tears streaming down his face. And we held an emergency lab meeting and resolved to pick up the pace of our research. So thankfully we're already supported by NHGRI. So by the next year we together with the Sequencing Center at WASHU had completed the sequence of the Y chromosome. And I think it's fair to say that it is purely and simply genomics that has saved the Y chromosome from the intellectual wrecking ball. So let me briefly summarize some of the things that we've learned having studied the sequence for some time now. First, give you a crash course, a little overview of the Y chromosome. So the centromere is over here and so we've got the short arm, the long arm. And I'm gonna emphasize, out at these green bits out in the ends of the so-called pseudo-autosomal regions where XY crossing over is a normal and frequent event in male meiosis. In between lies the portion of the Y chromosome that differs from all other nuclear chromosomes. In two respects, it's specific to one sex and it's the only part of the nuclear genome that does not ordinarily participate in crossing over with a homologue. The distal long arm of the Y chromosome is consists of rather simple monotonous repeats, heterochromatic and sadly no molecular geneticist has entered the heterochromatic region of the Y and returned alive. This I'm sure is why we as a community have steered clear in all of our sequencing efforts, steered clear of centromeres in heterochromatic regions. That was actually intended to be a joke but we actually have some challenges ahead and I will readily admit that though I just said we sequenced the Y chromosome, we haven't even touched this half of the Y chromosome that's heterochromatic. It's just emblematic of what we have yet to, what the challenges lie ahead in the rest of the genome. So we've instead focused on the euchromatic portion of the Y chromosome, which comprises about 1% of the human genome. And I will very quickly summarize some of the largest changes in our thinking sort of then and now about the Y chromosome. Well, so the notion was the Y was a genetic wasteland. We now understand it encodes at least 27 distinct proteins. There's a larger number of protein coding genes because many of these proteins are encoded by families with virtually identical members and there's a clear specialization in sperm production of most of the genes on the Y chromosome. The view that the Y was merely a rotting copy of an ancient autosome. We've learned that many genes during the evolution of the Y chromosome have been imported from the X chromosome or from autosomes, especially during primate evolution and many of these have been amplified once arrived at the Y chromosome. So much of the Y chromosome's gene count is relatively recent invention. I've been told for years that the Y is nothing but junkie repeats. What I will show you is that the Y chromosome carries features gene-rich palindromes of unprecedented scale and precision. The notion that the Y does not participate in crossing over doesn't have productive recombination was the explanation for years as to why all the genes must be disintegrating. What we've actually found is that an alternative form of recombination within the Y gene conversion probably sustains many of the genes of the Y chromosome across evolutionary periods. And yes, the Y is said to be headed for extinction. Well, it turns out, thank you very much, that even single copy genes on the Y look to be preserved through natural selection even in the absence of sexual recombination. And finally, the greatest of medical insults, of insults in this age of focus on translation, that the Y would have no medical significance, it's clear now that actually deletions on the Y chromosome are the most common known genetic cause of male infertility. This is from adagenic failure. It's also been implicated in testis cancer and as I will point out to you today, may actually play, believe it or not, a central role in the etiology of Turner syndrome. So I will focus now the rest of my remarks on the gene-rich palindromes and a possible connection to Turner syndrome. Okay, so let me tell you briefly about the structure of the Y chromosome's palindromes. You all know that palindromes are sequences of letters that read the same backwards and forwards, like madam, I'm adam. These palindromes on the Y are much, much larger than those you'd encounter in literature. They show very amazing degrees of left arm to right arm identity, as high as 99.997% or conversely, they show less than 0.06% divergence. I already wanna begin to plant the seed in your mind. That is less than the difference between alleles in the human population at autosomal and X-linked loci. These are big. The biggest of palindromes in the human Y chromosome has an arm length of almost one and a half megabases or a wing span of three megabases. That one palindrome is 1,000th of the human genome. And these aren't just pretty structured, they actually carry genes. The great majority of the genes required for somatogenesis in the human Y chromosome are born on palindromes. And this means that the one thing that we knew before any of this work, the one thing that we knew was the central biological reality of the Y chromosome was that its genes did not have partners on a homo. But what we came to realize was the genes on the Y chromosome carry their partners on the opposite arm of a palindrome, where they exist as mere image copies. And it turns out that these palindromes, and there are about eight of them in total on the human Y chromosome, they comprise 25% of the business part of the Y. And they carry all of the intact copies of the long arms test of specific gene families. Over on the left I list eight gene families and on the shown across are the doublets or couplets, the dyads of genes that occur in palindromic locations. And it turns out that knowledge of this whole of mirrors that is the Y chromosome has led to many predictions. It turns out even if that structure is very complicated, if you know it and you know a little bit about the principles of homologous recombination, you can predict variation and you can predict mutations and even predict in some sense what their phenotypic consequences might be. Many such predictions over the last 10 years have been fulfilled. For example, there are now six known highly reproducible deletions that occur de novo, but repeatedly and those are collectively the most common known genetic cause for metagenic failure in our species. What I wanna do is tell you about one particularly outrageous prediction based on the knowledge of the sequence of the palindromes and some of the simple principles of homologous recombination. So we made this speculative prediction that palindrome-palindrome crossing over might occasionally produce a mirror image Y chromosome, and isodicentric Y. So what if two copies of the Y chromosome, two sister chromatids, for example, were to align in opposite orientation? Palindromes would be equally fine templates for crossing over in either orientation. If a crossover occurred, you might produce a mirror image chromosome. Well, this might sound a little fantastic, but it turns out to be the case. So here I'll just show you that by in-situ hybridization with a probe, actually the red probe is for SRY, the sex determining gene, the green probe, for the centromere, if this is a pattern you'd see on a normal Y chromosome, this is what you see on some isodicentric Ys, and this is what you see on bigger isodicentric Ys, which had as their target for crossing over a different palindrome. So I'm summarizing here a re-examined study that carried out, just published a year ago, a year and a half ago, which we re-examined the Y chromosomes of over 2,000 patients who we'd studied because they had a discordance between sex chromosome constitution and their anatomy, or they had a microscopically detectable anomaly of the Y, or they were men with little or no sperm production. And out of those, we identified 60 individuals who had isodicentric or isoY chromosomes that are risen by the mechanism I've just shown you. 51 of these, I should say, arose through crossing over the palindromes, and essentially all of the palindromes, each of the palindromes that we knew about on the Y chromosome was hit as a target in one or more of these individuals. This turns out to be a significant cause of spermatogenic failure in our species. And then I'd like to just point out to you, sort of go in a little bit more depth into one particularly interesting aspect of the correlation with the phenotype. And that is that 20 of these individuals who carried mere image Y chromosomes were anatomically female. So I want you to think about how that could be the case. Well, perhaps some of these isochromosomes had deleted or lacked SRY, the sex determining gene. Well, that turns out to explain two of these cases. SRY is not present on the isodicentric. But that leads 18 cases where SRY is actually present not once, but on both ends of the isodicentric. So again, appealing to systems, to studies from, for example, yeast, became up with a hypothesis. And that is that these chromosomes, with two copies of SRY, they have two centromeres. This is not a happy formula for going into mitosis. The mitotic instability of these isodicentric Y chromosomes could have given rise in parts of the developing embryo, specifically to the existence of XO cells in the embryonic gonad. And this could lead to feminization of the external genitalia, anatomic feminization, despite the presence in skin and blood of the isodicentric Y chromosome. This led to a prediction. Turns out that farther these centromeres are apart, the greater the distance between the centromeres, the greater, predictably, the mitotic instability, the greater the probability of anatomic feminization. This turned out to be the case. So simply plotted for the 60 individuals, here are plotted the 40 individuals who are anatomically male out of the 60. And here, each of these dots represents the location of the targeted palindrome. And here are the 18 anatomically feminized individuals, which tended, as you see, to target more distal palindromes. Just to translate that, that translates into inter-centromeric distances, distances between the centromeres being, actually on average, about twice the inter-centromeric distance in the feminized 18 as in the masculinized 40. Fitting the prediction. So pointing to the existence, the very likely existence of XO cells. Well, to take this even further, so I should say in human beings with isodicentric Y chromosomes, the more Y DNA you have, the greater the likelihood that you will be a female. Led to further speculation. This is where I get to Turner syndrome. Could mitotic instability of isodicentric Y chromosomes in Zygos be a significant cause of the XO state in girls and women with Turner syndrome? Now, this should sound crazy to you, right? XO, isn't that due to mitotic non-disjunction? No, it turns out there is no maternal age effect in XO Turner syndrome, unlike trisomy 21. In other words, the mothers of girls with Turner syndrome do not have elevated ages. And it turns out that if you look at who contributed the X chromosome that XO girls retain, in three quarters of cases, the X chromosome is of maternal origin, which is to say the missing sex chromosome is dad sex chromosome. Now, since they're female, you would sort of immediately leap to the conclusion it's dad's X that's missing, but it could easily be dad's Y or an isodicentric Y. So I would like to suggest that Y-palindrome recombination could be a very significant cause of XO Turner syndrome. And I should say that none of this would we have been able to make any sense of, but for the availability of a very precise, highly accurate sequence across the most structurally complex parts of the Y chromosome. So thank you NSHRI for helping make that possible. Well, so how was the sequencing done anyway? Well, first of all, it required the dedicated efforts of not surprisingly some women. And I'd like to particularly highlight the key, the really passionate contributions of Tamoko and Helen here. Some men contributed to the effort, among others, Bob Waterston then at Wash U and with the work continuing with Rick Wilson there. Well, how was it done? I won't go through the details, but I'll just highlight a couple of principles which I think have more general applicability. So how would you sequence a large palindrome whose arms are more similar than alleles? Well, it turns out then that alleolic differences, if you were to choose clones from two or more men, alleolic differences would absolutely confound assembly. So you avoid alleolic differences by sequencing one and only one chromosome. Sequence just one man's Y chromosome. Now, of course, we didn't know that there were gonna be palindromes when we started this project, but thinking there might be some trouble ahead, we chose to sequence one man's Y chromosome. And it turns out that in this case, the left versus the right arms differ by only occasional nucleotide substitutions. So it actually worked your way across. The only way to do it is to capture a few such differences in large insert clones, like backs, derived from one man, so that's what we did. And then you use these left versus right nucleotide substitutions as unique markers. They become your landmarks. These occasional nucleotide substitutions become the landmarks with which you pursue the growing of back contigs. So you find additional such markers as you iterate. And so I would just call this, in some sense, haploid iterative mapping and sequencing. Just to give you an example. So here are two Y chromosome backs from one man. And they're typically sized backs. These come from the famous Mr. RP-11 on whom most of the reference sequence of the genome was constructed. And you will see this was a particularly inefficient use of NHGRI funds because these two backs overlap by 105 kilobases. We should have been taken out and shot at that point. But it turned out that within these 105 kilobases of overlap, there are 11 nucleotide substitutions. Or 99.99% identity. Now, if these backs had come from two different men, you'd say, oh, this is a lealic difference. In fact, it's barely a lealic difference. But it turns out these two backs come from one man. And these two backs turn out to seed where the initial seeding of the two arms of a palindrome on the Y chromosome. And so on. Well, we know that such structurally complex regions exist elsewhere in the genome. And that was recognized in Eric Green and Colleague's article just published yesterday, where they said that structurally complex genomic regions, which are known to have, and I might add the word disproportionate role in human disease, remain inherently difficult to sequence even with the new DNA sequencing technologies. Went on to say additional technological improvements, for example, much longer read lengths are needed to sequence such complex regions. I couldn't agree more. New technologies, which I hope actually appear on the scene, and maybe you all tell me that they're just around the corner, I agree that longer read lengths would be a tremendous asset in the pursuit, particularly of those naughty private parts of the genome required for replicating it, like centromeres and telomeres, that we're not even prepared to talk about entering yet. I think those longer reads will be absolutely needed there, but in addition, there's something else that I've already mentioned that's needed. So one of the key impediments in resolving the complexity of these regions is the diploid and polymorphic nature of the human genome, what I might call the allelic limit. In the past, the distinction between allelic versus polymorphic variation has been successfully circumvented by the use of genetic material of haploid complexity. This is not from the paper just published yesterday. The final sequence and assembly of the Y was achieved in large part due to the fact that all the back clones came from one man's Y chromosome. And so sequence assembly was not impaired by polymorphism and all sequence variants represented distinct copies of parallels. Now, this was actually written in 2002, not by me, but by Evan Eichler and colleagues in proposing the use of some unusual human conceptions that end up having a haploid representation of the human genome. And in fact, NHGRI made a considerable investment in resources based on these hi-to-tidiform moles. And I would like to suggest that the opportunity is right to take advantage of those resources, made some years back now with NHGRI resources, with NHGRI funding, in combination with the newest sequencing technologies, to extend the notion of haploid iterative mapping and sequencing to the approximately 160 structurally complex eukromatic sites on the X and the autosomes that are sort of crying out for a little closer scrutiny. I think, again, that the investment will be returned many times over. And so with that, I'd like to stop and thank you very much for your attention. Maybe we have time for questions. Questions for David? I'll ask one. So do your knowledge, has anybody taken a hi-to-tidiform mole, taken its genome, and just try to hold genome shotgun using the new technology? I think that is in the works as we speak, to give it a good look. But your proposal would be to take back clones that have been derived from those and do the painstaking work of developing the maps across these structurally complex regions. To go right to these regions, which I think are actually pretty much identifiable today. But not yet mapped. But not yet mapped in this way. And admittedly, the technologies for making those maps have an advance substantially in the last five, eight years. So it would be still pretty painful to make. I think it would actually be far more efficient today because using sort of back pooling strategies. I don't think it would actually be that hard. I don't think it would be that tremendously difficult. Yeah, Max, one more question. David, I have a quick question regarding the Y chromosome heterochromatin. I remember when we did chromosomes of medical students, the males, the medical students always seemed to like it when they had longer Y chromosomes than the ones with the shorter Y chromosomes who are a little jealous. But what I really wonder is what do you think, what's the difference? What do you think, what could these repeat sequences really mean? Because I don't think they're just junk. Yeah, males always like to say that mine is bigger than yours when it comes to the Y chromosome. And so there is, there actually, Max, it's a great question. One of the, ironically, actually if we go back now, 50 years, one of the first documented polymorphisms in the human genome was variation in the size of the Y chromosome. This was a very hot topic of publication in about 1959. And it turns out there is, to my knowledge, no known phenotypic correlate of variation in the size of the Y chromosome. And in fact, there are men who are essentially missing the heterochromatic region, who appear perfectly healthy, including their fertility. So the way I think about it, the Y chromosome minus that heterochromatic region would be by far the smallest chromosome. With the heterochromatic region, it joins the smaller autosomes in size. So maybe it's the kite that, maybe it's the tail that helps that kite sail better. I don't know, but not a very satisfactory answer, but thank you for asking. Okay, we're gonna move on. Thank you, David. Okay. Okay.