Transgenerational Epigenetics: of plants and men
This is a modal window.
Das Video konnte nicht geladen werden, da entweder ein Server- oder Netzwerkfehler auftrat oder das Format nicht unterstützt wird.
Formale Metadaten
| Titel |
| |
| Serientitel | ||
| Anzahl der Teile | 4 | |
| Autor | ||
| Mitwirkende | ||
| Lizenz | CC-Namensnennung 3.0 Unported: Sie dürfen das Werk bzw. den Inhalt zu jedem legalen Zweck nutzen, verändern und in unveränderter oder veränderter Form vervielfältigen, verbreiten und öffentlich zugänglich machen, sofern Sie den Namen des Autors/Rechteinhabers in der von ihm festgelegten Weise nennen. | |
| Identifikatoren | 10.5446/51098 (DOI) | |
| Herausgeber | ||
| Erscheinungsjahr | ||
| Sprache |
Inhaltliche Metadaten
| Fachgebiet | ||
| Genre | ||
| Abstract |
|
00:00
Vorlesung/Konferenz
06:09
Vorlesung/Konferenz
12:19
ComputeranimationVorlesung/Konferenz
20:28
Vorlesung/Konferenz
28:36
Vorlesung/Konferenz
36:45
Vorlesung/Konferenz
44:53
ComputeranimationVorlesung/Konferenz
53:07
Vorlesung/Konferenz
01:01:20
Vorlesung/Konferenz
01:09:34
Vorlesung/Konferenz
01:17:48
ComputeranimationVorlesung/Konferenz
Transkript: Englisch(automatisch erzeugt)
00:13
So, welcome everybody to this first seminar of the series of fundamental ideas on amazing logic of molecular biology.
00:25
And all these seminars are linked to a book that we plan to write with Misha Gomoff and Francois Kepes, which are also here. And all these seminars will be... So this book is about a brilliant and breakthrough ideas of biology.
00:44
About this seminar, so our link to this book will be about really outstanding and interesting topics in molecular biology and they are for all scientists. So the audience will be mathematicians, biologists, physicists.
01:02
And so today we are pleased to welcome Vincent Collot. So Vincent Collot is CNRS Research Director at IBENES, which is a biology institute inside RNS. And his main work interest, topic interest,
01:24
is epigenomics and epigenetics in Arabidopsis Italiana and especially transgenerational epigenetics and that is the topic of today. Vincent Collot, thank you. Thank you Jérôme and thank you Francois and Misha for your very kind invitation.
01:43
So I think I'll build up on what Anique presented to you just last week. I haven't listened to all of her seminars, but took bits and pieces from what she said and therefore it would be fairly straightforward for me to get to this topic
02:00
without too much introduction again. So yes, I'm going to talk about what we call transgenerational epigenetics. So what matters here is whether or not epigenetic states, and we will define them as we go along, can be transmitted across multiple generations. And not just actually between parents and offspring, but across multiple generations.
02:22
And therefore whether or not they can fall into the realm of genetics, which is a science of heredity. Okay, let's start with chromatin. You already had an introduction about that, so you have to remember here, or you don't have to remember, but I'm just going to talk about epigenetic phenomena that can cross generations in eukaryotes.
02:46
And most of what we know in terms of epigenetics in eukaryotes, whether or not it crosses generations or it's part of developmental programs, lies on the fact that DNA is not naked in the cell,
03:00
but is first within the confines of the nucleus and is organized in a very complex structure called the chromatin, which is made up of a basic unit, the nucleosome. And this unit is there to compact DNA, but to do just more than compact DNA, is to modulate accessibility of DNA within the confines of the nucleus within a eukaryotic cell.
03:24
So here you have a eukaryotic cell with a nucleus inside, and this is a very simplified cartoon depicting the DNA double-headed to start with, sort of folding into a more and more complex structure with a basic unit of chromatin, the nucleosome, which is made up of DNA of course,
03:41
and histones on which the DNA is wrapped. And then higher organization structures, which I'm not going to get into today. So what you have to realize is that we know that chromatin, sorry, we have known for a long time now that chromatin comes in many different flavors. And the first description of this was due to this botanist Emile Hartz in 1928.
04:06
So sort of the basic tools of microscopy, what he could find is that within the nucleus of this time most cells, you could distinguish two types of chromatin. Eukromatin, sort of a loosely packed structure here,
04:25
and what he called heterochromatin. So these two terms actually is terminology. And so this is the first time that you will hear about eukromatin and heterochromatin. And he also realized that genes, unit that will sort of dictate certain phenotypes,
04:43
resided mainly within eukromatin. And that heterochromatin was this sort of compact chromatin state that did not seem to change much from cell division to cell division, and did not seem to contain much genetic information. How do people do that at that time?
05:01
So very, very careful observation. And very good deductive skills, I guess. And some genetics. So the combination of the three, genetics... How could say which gene were in which part of the chromatin? The genes were not defined... Well, I invite you to actually read this paper,
05:21
which you can find already online. I haven't done this for a while, so I could not provide you an answer. But genes were not identified actually. The genes were abstract entities at that time. Absolutely. And I haven't talked about anything concrete here. No, no, say gene. But gene were not related to the chromatin at that time. It was not understood. Well, no, no.
05:40
Genes were related to chromatin in as much as we knew, I mean, from the work of Morgan, that there was this... There were genes, but they were not related to the molecular DNA. Nobody knew what was... I haven't talked about DNA here, yet. OK. So, gene is a concept. And that's actually...
06:01
genes that were proposed or invented more or less at the same time, one in 1909 and the other one in 1917. But what we have since 1903 is the chromatomatic theory of erudity. So we knew that somehow genetic information, whatever it meant, was carried by chromosomes. And then, you know, you add another 20 years,
06:22
you get some beautiful microscopy, and then people realize that chromatin is a substance of chromosomes. And genes reside within chromosomes. I'm not talking about DNA. Sorry, I apologize. I did talk about DNA in a previous talk. That's why you have confused. Yeah, yeah. I'm doing this introduction in the wrong way.
06:41
But I want... I mean, this is what you got last week, OK? What I want you to realize is that we knew very early on, before, indeed, we knew that DNA was the material... Yeah, OK. ...substance of erudity. We knew long before that chromatin came in different flavors and these different flavors were associated with different potentials
07:02
of expressing traits via genes. Genes being a concept, not a physical entity at that time. And, again, still with a historical perspective, built on these sort of early observations and subsequent observations, Marilyn Island proposed in 74 that different chromatin states,
07:21
namely eukromatin and ametrochromatin, could be associated with different transcription states. So now we have DNA. We have a central dogma established from DNA to RNA to protein. And now we bring back chromatin and DNA function because for a long time chromatin was only thought as a packaging device,
07:42
only there to put these two meters of human DNA into the confines of the nucleus, which is about 10 micrometer in diameter. OK. So chromatin was studied on its own, in the amount of gene function or the way genes were regulated. And Marilyn was one of the first few people,
08:02
I mean, together with Baba and MacLintock and others, to bring the two together. Actually, it's MacLintock who started that in a way. Well, I think that depending on chromatin states, you could have different phenotypes being expressed at the surface of maize kernels. OK. And here what you see is she also came up with this notion
08:22
of constitutive heterochromatin, which correspond to the heterochromatin that immune health defines, which is constantly condensed throughout the cell cycle, so not only during mitosis, but also during what we call the interface, when typically chromatin sort of unwraps a little bit
08:43
so that DNA gets more accessible to the transcription machinery, et cetera, et cetera. So, but this is constitutive heterochromatin and basically what she had is this coding chromatin, the one with function, which could exist in an active or in an active state. And what she had in mind here was the notion that even active chromatin could sometimes transiently exist
09:04
in an heterochromatin state. So you have constitutive heterochromatin and you have facultative heterochromatin. So there was a dimension of dynamics. Before these immunohides, we just had two types of chromatin. Now, with mammalian, we go one step beyond
09:21
and even open chromatin could exist in different states, and these states could relate to different ways genes are turned on and turned off. So I think this is important. I mean, it's a very, of course, a coarse way of describing what happened over the course of 20 or 30 years, but nonetheless, this is probably one of the early indication
09:42
that chromatin states per se could be very important in dictating or in contributing to different states of gene activity. And now this is a picture we all know. Okay, so we have chromatin. So basically, it's a nucleosome, and Anik mentioned to you last week
10:02
that at the end of Eastern, especially Eastern H3 and Eastern H4, we have all kind of modification that can happen on specific amino acid residues and these modifications somehow together color or put a flavor to the nucleosome that will be either carmesic for transcription
10:22
or repressive for gene activity. So you could modulate gene activity through this set of modifications. And there was also a modification on the DNA itself, not just on the histones, but on which the DNA is wrapped, but also on the DNA itself. And this is 5-meter cytosine, which we find in many eukaryotes.
10:41
Not in all eukaryotes, I'll go back to that in a minute, but in many eukaryotes. And in addition, I don't know if Anik mentioned it to you, but we also have what we call Eastern variants. Eastones are among the most conserved proteins in the eukaryotic world. And this unit is universal among eukaryotes. And yet, you do find a whole variety of Eastern variants
11:03
which replace the canonical Eastones, and when they come in, instead of a canonical Eastern, of course, they bring with them a whole new set of properties to nucleosomes and therefore to chromatin, which you could relate. It happens that none can do it like that. Very, very dynamic.
11:20
It's extremely dynamic. So you have chaperones bringing them in, picking them out, deciding, here, I want this one, this one here. Please, don't mess with me. But they bring you the same genes and modify? No, no, these are variants. So these are coded by different genes. Like different genes. So you have the canonical Eastones, which are extremely well conserved in eukaryotes.
11:41
And then you have these variants. And some of them are also very conserved. But others can now diverge between different eukaryotic branches. And therefore, this led to the notion of one genome. So we all have only one genome from our parents. So one chromosome set from our mother, one chromosome set from our father. And this is what we have to deal with.
12:02
And then, you know, to unravel this genetic program, we have to turn on, turn off different genes at different times. And this is what connecting will allow the cell to do. And therefore, for one genome, you'll have a multiplicity of epigenomes. Okay, so you're all familiar with this notion,
12:23
specifically what most people refer to when they mean epigenetics. Nowadays, it's basically different flavors of chromatin associated, causing consequence, whatever, of gene activity, of differential gene activity.
12:44
Now, among all of these modifications, one is particularly interesting to consider because of its inherent properties to be serving not only as a chromatin map, but indeed as an epigenetic map which could be transmitted across cell divisions
13:00
with extreme fluidity. And 5-8 cytosines. So this is the addition of the methyl group on the carbon-5 of one of the four bases that make up the DNA cytosines. And so we have known for quite a while in the 60s that indeed in many eukaryotes,
13:21
in particular in mammals, we have this modification as part of the DNA make-up of a cell. And two people, Omina Hadidek, and Art Riggs, in the same time in 1975,
13:42
both proposed that the mutation could serve as the toy-bitalating model that would convert different states of gene activity to a cell. And that could be faithfully transmitted across cell divisions, just like DNA can be transmitted
14:02
faithfully through replication. So here we have a similar system. And the key points about their proposal are that first, that of course, the mutation could affect gene expression. That changes in the mutation could explain the switching on and off of genes during development. So it was a developmental process,
14:22
a dynamic process. The predicated existence of enzymes that could mutilate DNA, first, you know, de novo, so to establish the mutation, and then a set of enzymes that could perpetuate these mutilation states for what they call maintenance mutilation activity. And that demutilation could indeed be irritable
14:42
thanks to the action of these maintenance DNA mutilations. And to cut a long story short, here is basically what we know about the perpetuation of demutilation states in mammals today. And here, there is one key component to this system,
15:01
which is that demutilation in mammals is almost exclusively found within a very specific sequence context, C-G dinucleotides. So when you read C-G on the top strand of the DNA, remember DNA is a double helix, so you have the complementary strand.
15:21
And what you have here is again C-G, so it's what we call a symmetric site. So when you replicate, so first, of course, you have to oppose demutilation in the first place, and you do that through what are called de novo demutilation activities. So you decide that you want to mutilate these two cytosines, for example,
15:42
for instance, in these two C-G dinucleotides, here and here. So you bring in demutilation, and now, just through the activity of this maintenance metatransferase, once you have replicated your DNA, you use information which is on the complex strand, this metallated cytosine on the complex strand.
16:03
And now what you realize is that what you have here is what we call a heavy-metallated C-G site, base pair. And now you just instruct somehow the metatransferase of maintenance to put demutilation opposite the metallated cytosine,
16:20
which is on the complex strand. So it is a semi-conservative practical process. It's not in detail how it works? Absolutely. We're not going to go there today, but the chemistry are beautifully understood. There is extrusion of cytosine, and we know that there is a key co-effector of this reaction,
16:41
which recognizes the heavy-metallated nature of this CpG-W piotile. Recognizes that there is only metallation on one of the two strands. And that somehow brings in this maintenance metatransferase, which is a different one from this one. This one is a de novo, in mammals it's called DNT3. This one is a maintenance, it's called DNTY in mammals. And DNTY only works with this
17:02
accessory factor on heavy-metallated DNA. So it can only do perpetuation. In fact, in vitro, this one can also be de novo. In vivo, its main function is to do perpetuation of metallation states through replication, through DNA replication. So this is a beautiful system. You first decide to metallate a gene,
17:21
for instance, to switch it on or off, and then you can perpetuate, so you can memorize the state across multiple cell divisions, because you have a system of metallates. And we know that this is key to many developmental processes, like X inactivation, or... Well, it's a bit vague there.
17:40
So once you metallate there, the system which does metallation become activated, because it causes your feedback in the system. There is no feedback. No, because you have to make a decision to metallate. Ah, you have to. This is a developmental decision. At the moment you metallate, you have a decision being enforced. Now feedback in the system, right around to it. Once you metallate, you block some gene.
18:00
Then again there is a process which does it, and this might be activated. So there are two decisions to make. Metallate and activate the process, which makes metallation. So it's another gene... No, this is... Okay, well, I guess you are right, but what we know is that this activity of denominator transfer is you can find it throughout development, in mammals, for instance. Ah, so it works everywhere.
18:21
Well, there are key, key moments, of course, in development, where you will find it most highly expressed and most active, which is during... in the germ lines, and let me get back to it, as well as during early embryogenesis. There you have a lot of denominator activity, but you can find
18:41
So I don't think we have a comprehensive picture right now of when this decision... No, but you're saying decide in particular place to decide. So it must be... Decision must be somewhere... Now, where... How do you decide? You want to regulate this... You have CPGs, but not the one next door. Right. That's another question. And we're not going
19:01
to go there today. I'm sorry. I can't do everything, otherwise you would hear me for two or three days in a row here. Okay. So now, this very simple system of memory, of cell memory, is critical for at least two processes in mammals. One is called X inactivation, and I'm concerned only two of us
19:22
in this room. These are the two female of the species that we have today. So basically, ladies, when you were conceived, soon after you were conceived, so basically, in a few cell divisions, there was a process which took place at random
19:42
in every cell of the embryo to inactivate one of your two X chromosomes. Because it's not good to have two X chromosomes when you only have one. You need to balance the dosage of genes that are sitting on X chromosomes between male and females.
20:01
We know that dosage is key to everything in development. If you have a trisomy 21, you know that you have phenotypic consequences. You have the down syndrome. What is trisomy 21? It's chromosome 21, smallest one of human chromosomes. Only 256 genes on chromosome 21. If you have three copies
20:21
instead of two, so 1.54 increase in copy number, you have the phenotypic consequences that we all know about. How long is the consequence that you won't initiate this chromosome? You'll never be here to ask a question. Basically, this is embryonic.
20:42
This is, I think the embryo dies at about, in the mouse within three or four days. Completely 30. Completely 30. You can't survive with two Xs. I'm sorry, with two active Xs. Of course, you have two Xs
21:00
in every cell of your body, including, of course, in your germline, because you want to transmit your X to your progeny, but yeah, only one of the two is active. And this is a decision we take very early, which is taken very early on during our biogenesis and to which you stick. Once you have decided in one cell to inactivate your X that you receive from your mother
21:21
or from your father, then all the progeny of that particular cell, so all these cell lineage now, we have kept the inactive X. Well, well, can we keep X from your father? Sorry? X may come from your father? It has to come from your mother. It has to. Yeah. The initial daughter is because you have
21:40
two Xs. We have two Xs. You only got one X, where did I come from? Yeah. Where does your X come from? Your dad or your mom? Yeah. My mother, I guess. You guess well. Absolutely. You only got your Y chromosome from your dad. She's quite observant. Yeah. No, that's fine. I made this mistake some kind of time,
22:01
and so once you have made this decision, this is it. You are stuck with it throughout your life, and that's why you get this calico, the cat. What is it? It's a cat, which is a female, and I'm going to use some genetic patterns, but, you know, you can't ignore it. It's heterozygous at this color code locus,
22:23
so one chromosome. I don't know if it was from the mother or the father, but one of the two chromosomes that this female cat inherited carries a black gene. The same gene here on the other chromosome now does not condition
22:41
black color, but orange color. Because of this phenomenon process, which takes place very early on of X inactivation, some cells will inactivate the black allele, and therefore you can only express the orange color code in this cell and all of its progeny.
23:01
Or you will inactivate the orange allele in another cell at random, it's one of the two, and of course now you are going to express the black allele. And I told you this is now something that is memorized completely. So all of the daughter cells from that cell to this decision will be orange.
23:21
All the daughter cells to this decision are now black. That's why at the end of the day when you have now a fully expanded cat, you have plaques of black fur next to plaques of orange fur and never the two together. Because this was a very early decision
23:41
in the blood count, and you have a sort of criminality then, sort of expanding. Yes? And you have the type you work out, the small one on there. Oh, this is DNA mutilation, my mistake. This is DNA mutilation, this key modification I was talking about. Black, mutilated.
24:01
Empty in circles, no mutilation on these cytosines, on these CPGs. Thank you. So this is it. And this is critical. As I said, you can't survive without it. Sorry, good question. Is the whole problem which is inactivated or is each gene... So it's a very important
24:20
question you ask. You know what, I think I'm going to do it as I always do when I talk about epigenetics. I have, of course, a program and then we're just going to talk. So yes, it's a... it's a very important question you ask. Is it a decision taken at the level of individual genes or the whole chromosome? It is the whole chromosome.
24:41
And it's a very beautiful and complex process. Again, I'm not going to go into any of the details today, but it is a chromosome decision which is taken by one particular locus in this chromosome. So it starts from one place but then it says, okay, I want my entire chromosome to be blanketed with whatever so that we turn it off.
25:01
But there are escapees. So in the mouse, not too many escapees between the two X's. So some genes will now still be expressed from both chromosomes. In humans, a lot of genes seem to be escapees. So there we have a kind of mosaic picture along the chromosome. Some of them will be subjected to this random
25:21
X inactivation process, which is a chromosome decision, but others will be sort of hidden from this decision. And we still be maintaining that, we mentioned the expression from both sides, from both chromosomes, sorry. Okay, so that's a really beautiful example. Another example that you probably heard
25:41
from Anik last week is genomic imprinting, parentan imprinting, as he mentioned that to you. The fact that basically, you know, we receive one chromosome set from our mother, one chromosome set from our father. We are diploid organisms. For each chromosome, we have a pair, in effect.
26:01
We have chromosome one, a pair, chromosome 21, a pair, XX, a pair of X, or XY if they are males. So we have these two contributions, but they are not equal either. You couldn't have the same DNA sequence, and this has been done experimentally in the mouse, and yet, some genes will only be expressed
26:20
according to the parent of your gene. They will be expressed if they came from the father, or other genes will only be expressed if they came, so this is the case where here, this particular gene called IGF-2 will be expressed only on the pattern really derived chromosome. It's not an exon. It's another chromosome. It's an autosome. Now, this particular gene
26:41
called H90, which is actually linked to IGF-2, now will only be expressed from the matinely derived chromosome. So now you have, again, a situation where the two are not completely present. We don't know still what it does, but the phenotypic consequences are quite dramatic.
27:00
You all have heard about mules and Henny. What is a mule and a Henny? It's the same genome at the end of the day, except that one was a male donkey with a female horse, whereas the other one was a male horse with a female donkey. At the end of the day, they still have one set of chromosomes from a donkey,
27:21
one from a horse, but the phenotypes are dramatically different, and this is co-producible. This is dictated by development, so this is a very robust system. There's only one gene in this case you're saying? No, no, no. It's plenty. I'm giving you an example now, but in the mouse I was about 100 loci that are subjecting to parenting and hunting,
27:41
and in humans it's not clear how many in a horse would not know, but again, so these two are really beautiful examples in wood, and they induce two examples. The decision is taken very young, sorry, the decision to be expressed or not is taken very early during development. In this case,
28:01
during early embryogenesis, here it's even more complicated. It's not within your generation that the decision might be taken, but in the generation of your parents. The gametes that the parents will transmit to make a new being, so the sperm and the oocyte, that's during gametogenesis and this is during oogenesis
28:22
that you decide to put delimitulation on whatever map that will now say, okay, this is coming from the father or this is coming from the mother. It will always happen for particular, I won't mean every beforehand, but the same gene or the same way? No, so it's different genes for different organisms. I'm sorry, no, for different sex. No, the same gene
28:41
for the same sex will go to the secretary. Oh yeah, so it's a rule. It's a rule. And it's mutilation also. No, it's not fun though. You always always mark this gene to be inactive when it comes from the female and this marking must be done during oogenesis. And always with mutilation. And it involves delimitilation.
29:01
That's a key point. Once you have that, then you maintain this information through the next generation. Of course, then you have both in both cases you erase this information for the subsequent generation because then you start with scratch. And that's what I want to illustrate here. We have this beautiful set of erasures
29:21
of delimitilation, re-establishment of delimitilation, erasure and re-establishment. This is a key property of mammalian development. This is the delimitilation level globally across the entire genome in male and in female. As you can see, both the oocyte and the spermatozoites,
29:42
or sperm cells, have high delimitilation content. But as soon as you fertilize the egg, you have a dramatic reduction of delimitilation. We believe that, for instance, of male clonucleus, this loss of delimitilation is an active process, an enzymatic process, where you actively remove delimitilation within one cell deletion.
30:01
So you haven't even started to do that actually. You fertilize your egg and somehow the heavily methylated sperm DNA becomes immediately delimitilated through an active enzymatic process, where you remove delimitilation. And then, amylogenesis proceeds, you get to this stage that's the system, which is a mass of cells,
30:21
and now you start this de novo delimitilation process. So you can expect that now DNMT3, this de novo delimitilation process, is extremely active here. And for whatever reason, this is so weird, but then you start again to lose delimitilation. This time you do it in the germline.
30:41
Male, of course, all the female germlines, you again lose most recognition, both sexes, and now you re-establish it according to your sex. So that's when you start re-establishing content. Okay? So this is very, very key.
31:01
We have delimitilation, providing memory, reset at every generation. Now, this is all well, and this is a kind of gene-centered view. The sheet is pretty hot. I know that it's raining, but I'm not going to survive for one hour or so. Leave it in the back.
31:26
So this is perfect. We are still in a gene-centric view of the world, and we want to understand gene regulation, and we have now found a function for delimitilation as a system to provide
31:40
memory of gene activity states. But one thing you have to realize is that now if you look at delimitilation across the human genome, first what you know is that genes is only a tiny fraction of our genome. How much of our genome encodes proteins?
32:00
How much? 1%. 1.5%. A bit too strange. 1.5. I give you 2% if you want. But that's it. Gosh, we are left with 98% that has absolutely no function that we can identify already. And what is most of this non-coding DNA,
32:22
as we call it? It's mostly transposant sequences. These sequences that jump around the genome, and just by virtue of this jumping, being able to jump around the genome, accumulate, accumulate, accumulate without having any function. At least we don't need to invoke any function. We just populate the genome. And the end result is that now half of our genome
32:40
derives from sequences that were originally retroviruses or other type of transposable elements. And that's illustrated in this particular gene, Parcatu, which is a gene involving, which when defective, leads to breast cancer. And what you can see that the exons are in black. Look how tiny
33:00
they are in the exons. The beta code for protein compared to all the rest. Which are now just a repository of transposable element sequences. Minions and objects, you know, they generate a community over time. But that's what we have in our genome. And basically, these sequences are the primary targets of demystification.
33:21
So when we talk about these big waves of demystification, remystification, demystification, remystification, it's mostly about rapid sequencing. Okay, genes are only a tiny fraction. They do also follow these waves. But we have to bear in mind that repeats are the primary targets of demystification
33:41
in mammals, but also in plants. And in mammals, it's beautiful here. I mean, this is really technical, but I don't know how familiar you are with the notion of southern blot analysis or northern blot analysis. We don't know anything about that. So southern is not because it's south. It's because it's head southern.
34:00
Okay, northern blot was a derivative of southern blotting. So northern is indeed the opposite of southern. But southern is someone's name. Anyway, so you can show that this DNA in white-type mice is ably methylated. I'm not going to do it in detail. But now in a mouse in which you deactivate
34:20
this maintenance genomic transfer as I talked about, dNMT1, now you lose demystification. Don't ask me how to read this, but trust me. And the same with now RNA when you do a northern blot to try to see expression of certain sequences So here we are probing with one of these endogenous retroviruses as we call them,
34:41
ERVs, plenty of them in human and mouse genomes. In a mouse, it's extremely active. IFP jumps like crazy. Now in a white-type mouse, IFP is completely sign-on, no expression, no transcription. But now in a dNMT1 mouse, you have removed demystification. And look, now you have this huge activity,
35:01
this huge transcription activity, again showing that demystification is a key component of the control of transposon activity in the cell. Okay, so this is... What happens to the mice? Dead. Sorry. I forgot to tell you that. That's obvious. Dead. But it dies at 9.5 days.
35:22
So, you know, it can go through a few... By first, the first division is okay. It's not immediate. Absolutely. Yeah, yeah. But why does the jumping is so bad at the beginning? It dies at 9.5. But why the jumping can go through genes is not so bad at the first stages of... I know it's bad, but, you know, there are ways of resetting, so these things
35:41
are stochastic events so they may jump in places where there is no other gene so that's okay. You know, it's kind of it's a kind of catastrophic event. But the rate of jumping comparable to the rate of division of the... So, actually, here we don't know about jumping. What we know is about transcription. So, these transposon sequences get transposon transcribed. So, of course, now they can jump. But we haven't actually monitored
36:00
transcription probably. Yeah, okay. But transcription can also be deleterious because these sequences that you remember, sorry, they are everywhere within a gene. So, now if you transcribe one of these, you know, you score the proper transcription of this particular gene. But this is actually, sorry, this is a human copy and this is a mouse copy.
36:20
Two different sides, but both are made up mostly of repeat sequences. Okay, I'm almost finished with my introduction. I want to tell you that we're bored. I do. So, I have to move on, I guess. I do have to move on because you know all about this.
36:40
No? I don't know. You said yes. What shall I say? Okay. I'm going to focus on pigmentation because this is really the topic we study in the lab and that we try to associate with this transcription of genetic. But I want to remind you that epigenetic mechanisms, even though they are
37:00
very widespread among eukaryotes, they're not universal. You find many instances of organisms which don't have delimitulation. Two yeasts that are used as lab systems, like cerevisiae or pombe, do not have delimitulation. Drosophila monogaster, the fly,
37:20
no delimitulation. C. elegans, the nematode, which is a beautiful genetic system to do genetics, of development, no delimitulation. So you can have life, eukaryotic life without delimituation. I've had my smaller genomes again. Is that what you're mentioning? Most of them, but I'm sure, I mean, your point is correct
37:41
that you do have actually organisms with moderate size genomes that do not have delimituation. So it's not really simply a question of quantity of parts with a very low amount of content. No, no. It's not as simple as that. There are other ways of controlling that control zone. And some of these other ways are this mechanism of RNA interference that you heard a lot about
38:01
from Anik last week. So this is a very important system which can go in parallel or in conjunction with delimitulation to control repeat sequences. Okay, so that's important when you hear around, you know, all this excitement about delimitulation and so on. Okay, we have it. Plants have it, but not everyone has it.
38:21
And there are ways of controlling these repeat sequences or memorizing a gene activity state that do not depend on delimituation. I'm just going to drop one word or one name. Here it's polycom. That's another way which is different from delimituation by which you can perpetrate gene activity states.
38:41
How old is the family? Sorry? It's the mouse. It's the mouse, yes. The mouse. That's all. Because people are obsessed with the mouse. We have a mouse one. So the mouse has many things. It has delimituation plus the smaller
39:03
many pathways and different pathways. Plants have both systems as well. And then you have all kind of variations between these two. Now, what we're going to do is the subject of my lecture today. It's genetics. And genetics starts
39:21
117 years ago now. It's a rediscovery of Mendel's laws of inheritance. And I'm going to describe them. You all know that. There is no blending. There is segregation. There is independent assortment of characters that are not
39:40
associated with each other. For instance, you know, the color of petals, the size of fruits could be separated and therefore they will have their own rules of inheritance. But what have we done? Since 1900 until we got to the Saint-Trois-Beaux-Mains we have equated genetic information
40:03
At first genes were just a concept. But in 1953 that was the end of it. Because now we knew how you could perpetuate genetic information encoded in the DNA from one cell division to the next and of course from one generation to the next through this semi-replicative,
40:21
semi-conceptive system of application. Yeah. So when you have an A you know that you have to put a T opposite. When you have a G you know that you put a C. So we have and everything that is genetic must be written in the DNA sequence. So if you have a white petal and a pink and a white petal
40:40
and you can show that they are breeding two so basically they are two different alleles of the same gene so one has a form and the other one has another form you have to find a DNA sequence difference that corresponds to these two different forms. That's what we do when we do human genetics nowadays. We want to associate
41:00
differences in DNA sequence between all of us with differences in predisposition to disease and et cetera, et cetera. All of these genome-wide assertion studies. Yeah. Well now this is also part of the reality not everything is actually described or written in a DNA sequence and this is
41:20
a beautiful example in tomatoes where you have a mutant phenotype. You have a foot here that cannot ripen so this tomato is called colorless non-ripening. This mutant breeds through so basically it follows all of Mendel's laws
41:41
the two characters sorry the two is very uncompared to the Y-type state does follow Mendelian neurons. It's a single locus with a recessive mutation but when you see transmers of DNA of CNR compared to the Y-type you end up with the
42:00
same DNA sequence. So you can do genetics genetics that's what we are doing Mendelian genetics of course this is an extreme case but I want to use this to make the point that we can do genetics without differences in DNA sequence what you have instead is difference in chromatin states in DNA
42:21
methylation states so here and we know quite a bit about this particular system here you have one locus which is responsible for mutant versus wild type so this has been determined genetically it's all kind of tricks that we have these markers on the chromosome so we know
42:41
that this is a locus which is responsible for this difference as I said when you sequence you get the same DNA sequence in both cases but what we find is that now this particular allele which is actually the same allele as this one is heavily methylated in its close to the gene whereas this one is not methylated
43:01
and this methylation associates with silent signal and this one is not we see the DNA difference in the somatic cell or in germ cells everywhere of course the studies have been done in somatic cells in germ cells we believe that this is not the case so here
43:21
what we have is what we call an epimutation a change in epigenetic state of this particular locus or this particular allele as the locus now as this allele which is the same as this one can exist in two alternative epigenetic
43:40
states what is remarkable is that this heavy methylation associated with silencing now is according to Mendel's so this is now a repeat
44:00
sequence a transposable element I want to remind you that this is the primary targets of the immunization in plants as well as in mammals so we believe that what happened is that you know you have this transposon sequence upstream of the gene and somehow of course it's so it is paramount and as a result the gene is perfectly
44:20
expressed here but now these mutations seem to have spread through the nearby promoter sequence of the gene through the nearby gene and this spreading now is resulting in this shutting down of the gene and therefore the mutant can attack so we also probably have a molecular
44:41
mechanism here to propose it's not a mutation in a sequence it's an epimutation a change in epigenetic states associated with a change in DNA methylation but now it's not just plants which do it mouse mice do it too
45:01
but it's more complex despite this programming that I talked about in methylation states you can have situations but again repeat sequences these are the transposons I always do them with a trimeter are inserted within a gene or next to a gene and depending on the methylation status of the transposon sequence now the
45:20
gene is on or off so you have the same DNA sequence but now the gene is on here the gene is off and somehow these different epiallelic states can be transmitted to the protein it's not purely mandelian it's more complicated now this beautiful sorting
45:40
of peptide colors or whatever or shape but it is still transmissible and it follows some kind of entrepreneurial pattern the same with this particular mutation here axiom field depending on the methylation status of the repeat sequence the gene will be on and off
46:00
and this state can be transmitted independently of any change in DNA sequence that's what we call transnational epigenetics changes in the chromatin state that do not correspond to a change in DNA sequence that can be transmitted through multiple generations
46:20
so now this is not controversial but what we know at this stage is how much of that occurs in nature so basically how much of what we call genetic variation nowadays when we think of genetic variation the fact that we all differ by certain characteristics which have been transmitted from our
46:40
parents how much of these genetic differences are actually caused not by changes in DNA sequence but changes in chromatin states that can be transmitted through multiple generations so now we're not in a developmental context now there is no resetting we have two alternative states or maybe a quadational
47:00
state as you can see that you exist in a multiple range of states between fully mutilated and with a yellow fur and fully mutilated with a agnotic fur but you can also be intermediate with some transmission of these intermediate states so how much of that genetic variation that we describe around us
47:21
is caused by similar things and that's developed the human genome in 2001 not just
47:42
because it's beautiful because we believed I mean not me but people who were promoting this believed that by having the entire sequence of the genome about this cause that really helped us understand the genetic origin of these things
48:01
because now we can monitor differences in the sequence between individuals and associate these differences between individuals two differences in phenotypes which could be narrative and there is one beautiful example of character phenotypes in humans height we all differ in height here not because of what we
48:20
ate when we were in our mother's womb or what we ate during our childhood but because of what our generation in Japan the average or even the judge the judge we can take we can factor all of these effects
48:41
and still at the end of the day you can explain what 80% of our differences in height here by what was transmitted in Japan you see by age in Japan it's by age you can see immediately a man please please believe me that is the confusion here globally
49:01
species level differences in height in humans are mostly determined by our genes and genes I'm not saying DNA genes so it's genetically determined and we have tools now to associate this humongous level of variation in
49:20
the sequence between the individuals with differences in height and the study has been performed starting with 10,000 individuals and a million polymorphisms across the genome and then 40,000 individuals and 10 millions of polymorphisms and when you combine all of these studies of the order of 20 to 30 million
49:40
euros you end up explaining 10 to 50 percent of the 80 percent direct ability that genetics predicts so the trait is genetic 80 percent of the variation between any of us here has to do with our genes not of our environment and yet when you look at
50:00
the DNA sequence to try to explain it using the power of whole genome sequencing and huge co-op etc you explain 10 percent of this 80 percent so what you have is this huge problem of missing irritability where is genetics gone yes
50:21
it's a genetic definition it's in quantitative genetics so I don't want to go into the details just again but this has been this is 30 or 40 years of mathematical modeling of genes multi-genes etc and how you accommodate for that away from environmental influences that's a key point so
50:40
yes we have this huge problem of missing irritability so what I'm going to tell you today is not that I have the answer for this problem missing irritability is for epigenetics that epigenetics could be one of the contributors of this missing irritability if we have the example of other phenotypes in
51:01
the mouse plus what we know in plants why should we discard this notion that differences in chromatin states could indeed be transmitted across generations and contribute to to to to part of these heritable phenotypes that we measure
51:23
okay how much do you want me to talk about you know I'm happy to stop here no no totally don't don't feel embarrassed you don't
51:45
care about my own work what you care about the concept I'm happy actually to discuss with you and to show what we have done okay in that shape what we have done is to show that this is not this is
52:01
not marginal we have ways now of measuring it of course in a other system which is plant you can't do it in humans you can't because of Shut up because it's too much we can manipulate your plants without compromising
52:20
our diary so we can do genetics which of 630 from exc teased can be transmitted across generations, independent of anything else happening in the genome. So we fix our genome and we simply play with what we call the metelome after the epigenome
52:46
and how much of that can be transmitted. And we found that even in a genome as simple as one of our galaxies where repeat sequences, which may be the mediators of these epigenetic effects, I'm talking about the one that go across generations, even in our galaxies where we have so
53:02
few repeat sequences, we can find thousands of loci called the telomeres, for which a chance in the telomeres that we have induced experimentally can persist for 10, 15, 20, 30 generations and follows Mandel's laws. So there is genetics, absolutely, there is a dimension of genetics which does occur
53:23
independently and in addition to the one contributed by DNA sequence variation. So that's my take on message. I'm sorry, I cannot convince you because I don't see if it's, I have time to show you the result. No, you don't need to know. Honestly. The defect. It's boring. Trust me.
54:07
You're right. You're right. I mean, and I give you part of the explanation. I didn't, I didn't draw you with more details, but I gave you part of the explanation. So we have these ways of demystification and remediation. So of course you don't expect epigenetic alteration that will happen by accident.
54:24
That's the one I described where you, you know, imitation spread from a repeat into a nearby sequence to be firstly transmitted across generations, see if you can erase it and then start to scratch. But despite this beautiful system of double erasure, right early during a biogenesis and
54:43
in the two germ lines, you still have this example I described. You still have agotibial yellow and that's infused and I can show you, I mean, you know, it's complicated because now the genetics is not truly, truly Mendelian genetics. So the transmission across generation is more complex and it is extremely influenced
55:04
by the environment. It is extremely influenced by the genetic makeup. So all kind of things that we can deal with much more readily in our experimental systems with plants. But there is no reason to believe that this does not, I mean, first it is what we don't know and I'm therefore going to conclude with my conclusion.
55:24
I'm sorry about all of this skipping, but you know, that's fine. That's fine. Honestly, what matters is this, that we can induce disability variation which is associated with transport and sequences, quite readily in the lab and I'm sure I know if we can't
55:46
do it in mammals because it's not viable. That's the only, but here to me this is, it's not because there is a fundamental difference between mammals and plants. It's just that in mammals, the manipulation has been now co-optive for gene regulation
56:01
to an extent which is much larger than in plants. The primary function is to target with its sequences in both systems, but now in mammals you have this selectin activation, you have this imprinting, which has taken quite a bit of space functionally. And you can't dispense with that. In France, we have a bit of genetic imprinting.
56:21
Somehow it's only found in one tissue called the endosperm which surrounds the siempre-yoke in the sea, but not elsewhere in the development. So it's inconsequential if you can survive, you know, sea formation then you can make a plant. Okay, so that's the first thing. The similar thing, I haven't told you, but what we have established in the lab, we
56:44
find in nature. We have the tools now in Arabidopsis to go to nature, collect thousands of accessions of strains across the globe and sequence their genome, not only sequence their genome, but also identify their metallome at a single cytosine resolution.
57:03
And what we find is that the EPID variation that we induce in the lab also exists in nature. And in the lab we can establish that this epigenetic variation that we induce has different stability. Depending on sequences, some of them will be transmitted for hundreds of generations,
57:23
but other sequences will revert back to the default metallated state. It's crazy, how could we? Why? What's the difference? Yeah, but that's too molecular. I mean, I don't think, I don't know. I mean, I would love to tell you another time. We have the mechanisms, that's to do with the way these sequences are targeted by another system that's reinforced in the methylation machine.
57:42
On the sequence surrounding the site, yeah, the methylation mechanism, all it knows in the possible reproduction, only surrounding the DNA, right? So it might be arranged by the sequence. So DNA, the sequence tells you already what? So there is absolutely a component of DNA sequence. So that's what we find when we create across the genome all the disability variation.
58:05
We realize that they are not all equally stable across generations. It's just getting important in DNA. Absolutely, you're right. It is including the DNA, ultimately. But now, what you have to consider that, okay, we have a sequence here, which has
58:20
some feature, for which we can play with gene regulation states across multiple generations. In a penalty of the rest, the rest is completely, I mean, sorry, we never touch the genome sequence. Yeah. We simply force demethylation, et cetera. And now we realize that this loss of methylation that we can use experimentally
58:42
over repeat sequences can be transmitted. We haven't changed the DNA sequence there or anywhere else in the genome. But now we have another sequence in the same genome, for which we also manage to lose, I mean, to induce a loss of methylation. And what we realize is that after two or three generations, methylations come back to it.
59:00
So there is very strong conversion to default methylated state. So what I'm saying is that this sequence and this sequence are different in their behavior and showing us are different when we look at the sequence itself. So there is a feature of the DNA sequence itself that is more permissive.
59:22
Over some repeat sequences? No, so, I mean, we have to do machine learning. We have to come to mathematicians, physicians, whatever. How long does this need to be? So it's a few kB's because these are repeat sequences. And so, unlike in there, but, you know, we can only do it by first
59:40
artificially inducing these changes in the lab in a constant genome environment. That's really a good idea because epigenetics, I told you, is the fact that you can exist in two different chromatin states or methylation states independently of the DNA sequence. And now you can transmit it as well independently of changes of DNA sequence. Now, different.
01:00:03
So clearly, there are things that are written in the DNA sequence itself that tells you, yes, I can be permutable, or no, I cannot be, because I will always get back with high efficiency, my default mutilated state. Now I see that this is what we establish in the lab with a constant genome.
01:00:22
But when we went to nature, we found exactly the same type of differences of mutilation states, and of course, the one that in the lab we define as the most stable ones. So a change in mutilation state that you can transmit for many, many generations. We find it more often in nature than the one for which in the lab we can only exist
01:00:41
in the non-default state, which is the unmutilated state for one or two generations, because immediately it recovers with mutilation. So these ones in the lab, where we find them in nature, because they always go back to the default state. So this place is more in the genes or in the- No, it always repeats. No, and each one of them repeats. All repeats, but actually no.
01:01:01
These repeats have an impact on nearby genes. So that's how it's working. It's not gene regulation. It's transposon control, but with a consequence on nearby genes at the key point. So there is an amount of genetics which is completely away from genes here,
01:01:21
completely away, but you cannot identify it with a change in a sequence or whatever, but which clearly contribute to heritable differences in phenotypes as well. So the last thing that we did in the lab, of course, is experimentally, molecularly demonstrate that we could have these different metabolic states generated from,
01:01:41
and transmitted from multiple generations or not, so that we could associate these differences, these differences in phenotypes. With precision, transposon may change, because this is in the mutilation, the first stage. And that's another thing that we, of course, explore, is that once we have transposons being manipulated in terms of mutilation states, as I showed you, we can reactivate them.
01:02:03
And if you reactivate transmission in your transposon, now you can start to jam. So we have this sort of, I mean, it's not a problem, but we have- In the example here, in the first stage, when they give mutilation. Yeah, so we can, yeah, and so we have, now we are playing with those. I mean, not playing, but we are taking into account the two aspects.
01:02:20
So purely epigenetic aspects, so a sequence can exist in two alternative states for different stabilities according to our chance. And that contributes, therefore, to genetics to a different degree, depending on the stability. And we have now these sequences that we have artificially moved into the, not moved, we have shifted to the unmutilated state, this non-default epigenetic state,
01:02:44
and that now start to be, every year, transcribed, and some of them start to jam. So now we create what people call genetic variation, what I would call DNA sequence variation. Both are genetic variation, because they follow the next rules, or rules of inheritance.
01:03:01
So now what we can do is see what of the two seem to generate more phenotypic diversity, which could be noted. Is it the jumping of transposons, or is it the alternative epigenetic states of the residing transposons? And clearly the two contribute to phenotypic variation.
01:03:24
And I think between humans, or mammals, and plants, there would be this issue of, again, of transposon mobilization compared to epigenetic variation in transposons, plus the stability across generations, which we predict to be far less in mammals than it is in plants,
01:03:45
but not completely ruled out. And if some of you really want to know, I mentioned this retrotransposon sequence, sorry, this transposon sequence I use, we have all this terminology, I apologize for that,
01:04:01
endogenous retroviruses, as those are called then, IAP in the mouse, extremely aggressive, it jumps like crazy. Well, IAP, when we see these waves of methylation, demethylation, in the early embryogenesis and in the genes, in general, IAP stays methylated all along.
01:04:23
So they are not reset. So now imagine if by accident you lose methylation over an IAP, and this has no mechanisms to reestablish it, because it does not go through this programming of loss and gain. It's supposed to be by default methylated, like we have in plants.
01:04:41
Repeats are supposed to be methylated, and we see them methylated throughout life. Okay, we don't see this at all. And now the IAPs behave like a plant repeat, concentrated in the young ones, the ones that are the most aggressive, and indeed are the ones that also created these APA alleles at which I have a yellow, and that's
01:05:01
essentially the two phenotypes that I showed you. Go ahead, please. So during this report, I mean, it never goes to zero.
01:05:21
Okay, so then the methylation, which is alpha, is a random? No. So it's, again, So what you're keeping information, you're keeping information up? No, so it's very interesting, actually. First, what should, yeah, so let's look,
01:05:41
I mean, now we're talking about mammals by definition, because we have these waves of methylation and methylation. By default, the entire genome is methylated. Every cytosine in a CpG nucleotide will be methylated, except when you have
01:06:00
a high density of CpGs that are the so-called CpG islands. For whatever reason, these things get protected from the methylation. But otherwise, you lose everywhere, and you put back everywhere. So there is no targeting to be done. But now, as I said, some small subset
01:06:20
of repeat sequences never go through this. They go maybe like that, like that, but that's it. So they're actually heavily methylated soil planet. So in those, those which are, which transmit genes in this state? Well, these are these active transposons sitting in genes, which,
01:06:41
for whatever reason, we still don't know molecularly why they lost methylation. But if they're not part of this programming, you know, if you lose it, then there is no instruction to remediate, maybe. Even though it's by default, you should meditate everywhere. How particular location for methylation can be inheritable?
01:07:00
If they in some stage are developing, they all disappear. How do you know where to come and where to meditate? That's what I'm saying. There is no, the default state, it's called the genome, it's for methylation to be, to be enforced. No, but then, but then it still is different, right? It's inheritable, features some region which is not this inheritable. So in some moment, it being lost, everything being meditated,
01:07:22
become exactly information getting lost completely. So if everybody is meditated in this region or demitrated, we don't know completely. So there is very little dynamics of the methylation in relation to gene regulation. I'm contradicting myself here.
01:07:40
What we have is this beautiful association between methylation states and exon activation and imprinting. But again, so what you have to realise is that now this is a few genes, well, a few genes. It's a thousand or so genes on the exo. Maybe a thousand genes that are involved there. So you all have this CpGI norm I'm talking about.
01:08:02
So these by default are never methylated, except if you decide to target them. But this is a minority of the methylation that you find in the genome. The majority of the methylation you find in the genome is of course in non-CpGI norm by definition and it's everywhere.
01:08:20
Then they cannot be inheritable. No feature of methylation can be inheritable. You can lose information. So you have some feature of the organism and then some material is not and then some of it is being information lost so you can't recover it. So it cannot be inheritable. The way I describe it, we are purely norm which will be a material.
01:08:41
So, let me try to, I think I understand your point or your question but let me try to rephrase what I said so far. We know that we have these ways of demethylation or methylation. So there clearly there is mechanisms to dictate first to remove methylation across the genome.
01:09:02
It's not a perfect process as I said. There are some sequences that escape from this and then to say, okay, let's remove it. But when you look across the genome it seems that most cytosines will be methylated. So there is no much targeting of specific sequences. You see what I mean?
01:09:20
So then it cannot be inheritable. Exactly. Yeah, yeah, it's not inheritable. Yeah, oh yeah, of course. And yet, yes, it's not inheritable. It's not inheritable for sure. Those things are, and that's why you, you know, things work in mammals because they are not inheritable. But you have examples in which you have now these aberrations. Ah, these are aberrations.
01:09:41
This is not inheritable, this is aberrations. These are aberrations. This is not a very healthy mouse actually. This one is diabetic. This is a healthy mouse. It's a mutant. Okay, so this is really rare. I mean, you know, we are all different in our susceptibility to diabetes. We are all different in height. These are genetic.
01:10:02
But this is a rare case. So this being inheritable is a rare case. Is that again what you said? Yes, these are rare cases. These are exceptions. But I don't know what exception, I mean. How exception are they? I have no idea. And could they be, you know, a big part of this missing irritability? I have no idea.
01:10:20
That's a question. So this wave may be not so uniform after all. It's not uniform. Exactly. My pleasure. Okay. More questions? Yeah, well, you can see we missed something. We missed a lot.
01:10:41
Yeah, exactly. I mean, there is no way we can, no. I mean, I'm not going to do another talk. No, no, I mean, that's fine. I think you're getting the message across that basically, at least in mammals, we have a system which will ensure that there is no such thing as transmission across generations
01:11:01
of these mutilation states. But you have accidents which have been documented, which are complicated to document because the genetics is more complex than the simple Mendelian genetics of, you know, peas, shea or petal color.
01:11:21
But nonetheless, there is familiar transmission. You know that there are pedigrees you can follow, et cetera. And it's not just cultural transmission. It's definitely, you know, the mouse you can play around and show that there is an amount of inheritance there. And then there was this issue
01:11:41
of not being able to explain much of what we call genetic inheritance. In humans. With your simple set of genetic environments that have been collected. But this variation, this 80 percent, the inherited ones or just random color? No, this is a
01:12:02
genetics is a key concept. So you expect to be able to document it with what you inherit from your parents. The genes. No, but maybe not the maturation, not the assumption, Exactly, that's what I'm saying. That could be something else. Some other variation, yes. Which are not inheritable.
01:12:21
They're just variation, how many of them do I have? No, no, no, no, no. They're inheritable. We have the tools to analyze these things. Absolutely. But you don't have so many generation of humans to say they're inheritable. Yes. You need three or four generations. You need three, you know, Say four generation of humans? You need two on the back on that side and you need three on the back on that side.
01:12:41
So mother has, of course, and she's pregnant, she has a fetus, and the fetus has its own germline. So you need to go one generation after to make sure that it's not an environmental effect. For the father, you just have to make sure that the father carries his comorbidity. So it's comorbidity. But you cannot see if you're not typically right when we're looking at the
01:13:00
germ cells yet. You have to actually observe the variation between humans. Non-variation in DNA, but you have to. Variation between humans you cannot see so easily, right? It takes time. Well, you have variation which is caused by the environment, and variation which is caused by through regression between parents and offspring, you can show that this is genetic.
01:13:21
So we have ways of calculating and formalizing these two contributions to... Which you don't know when you park in an environment. Most of them you don't know. That's fine. It's a big error. I mean, that's what you put into your environmental factor. It's only the things that you cannot account for through pedicure analysis. Okay, so the environmental factor is 80% is environmental.
01:13:43
80% is non-environmental. But how can you say that? Not me. These guys. 60 years of genetics, of human genetics. I mean, you have to teach them maybe how to do proper genetic studies, but the consensus now is that... Yes, I don't know
01:14:01
what is the base word. It's quantitative genetics. So you have ways of partitioning, through pedicure analysis, what you can... You're measuring phenotype in so way, how can you say whether phenotype is similar or not, what's future present or not present. Well, you know, height is a quantitative trait,
01:14:21
so you have... It's exactly about height? Yes, this is height. All this about height? Yes, but I mean... So the principle of... I'm sorry, most genetic traits in humans are quantitative traits. Disease, if we focus on disease, diseases that are caused by a single
01:14:41
gene being defective are minority. Most diseases are caused by what are polygenic, polygenic in the sense of the term. So are caused by many variants, over many, many genes. So we are now... No, but also this B.O.M. also is made from stromatin steroid, yeah. B.O.M, you have bacteria
01:15:00
living in the gastric group, have a contribution of half of that, right? Yes. So you have to... So you don't transmit your... Which is partly inheritable. Partly, but not completely. So I mean, all of this is being factored into this equation that allows you at the end to determine that the whole ethical sense of differences in human height
01:15:21
is genetic. Of course, I don't know what percentage means in this sense, yeah. It means that if you measure differences between all of us here and it has, of course, to be large populations and you take care of the history of where you were born
01:15:40
and therefore the kind of food that you were exposed to, whether it was famine, in one generation or two generations. Okay. So you put all of these factors into your... And you still have to become dominant, yeah. A mixed genome or whatever. At the end of the day, you still end up with 80% of this variation that you can attribute to pedigree.
01:16:02
It says it's genomic or epigenomic now. So it says... Genomic. It's purely gene. Genome. That's what these guys mean. That's what the premises of my presentation. Then it says... Everything according, you know, genetics nowadays is equated to DNA sequence variation.
01:16:21
And what I'm telling you is that this cannot be the case. But only 5% is understood. I got confused about this picture. So you know 80% genetic, but then the second diagram says it's 75% missing. Yeah. It's genome. It's missing in relation to DNA sequence variation. A correlation between genomic
01:16:41
sequence and... Exactly. This correlation can only explain 5%. Alright. Here it was 5%. So I've improved it now by increasing the... Wait, wait. You know it's determined by genetic, but what exactly you're missing? So you know sequence, forget about this 20%. Sequence determines your height, so to speak. And then what you're missing?
01:17:00
Well, that's what I'm saying. If you think that sequence is determined by your height, you are falling short of information because if you take all the information in a sequence that we have now, then we have a lot. We have a coefficient of individuals that have been sequenced. We know all their variants around the DNA. And when you do, you're at sufficient correlation
01:17:21
with height. But you don't know which part of the genome, because since you have huge space, you just don't know what to correlate. So it's a kind of logistical... Well, you say it's additive. This is what we call additive. It's not biological, kind of combinatorial. That's such a huge thing that you really don't know what it is. Well, that's exactly... I mean, you know, these are very important questions that we have now,
01:17:41
that we are facing in the field of conscious genetics and human population genetics in particular. What people realize is that clearly, they want it wrong. It's not such thing as 80% collectability. Maybe it's much less. Or they have to look for other explanation than this variant they are
01:18:01
looking for in the DNA sequence. We know that most of the variants that have been used in these studies have what we call SNPs, single nucleotide polymorphisms. It's only part of the variantation. We have copy number variants. We have rare variants with large effects,
01:18:21
so-called rare alleles with large effects. They're never computed into this calculation. Ah, I see. Only SNPs, anyway. Only SNPs. And the SNPs are clearly only part of the explanation. And that's the problem we have. And, you know, now the latest came out
01:18:40
one or two weeks ago, if it was in science or in nature or whatever, there was an opinion on this and saying that, well, maybe every bit of, every sequence, every nucleotide in your DNA contributes to your phenotype. And therefore, it's not just maybe 10, 20,
01:19:01
30 major genes that are important in determining height. Maybe it's almost the only thing which defeats the purpose, such genetic origin of disease or whatever. You see what I mean? So they are talking about only genetic disease. Of course, it's a surrender.
01:19:22
I mean, maybe it's a reality, but then, you know, forget about any of the approaches we have been promoting for the past 20 years, which is to try to find the set of genes of flow size, assuming that there are not just one, but not thousands, that determines
01:19:41
these major differences between us. So in two minutes, imagine that there are five nucleotides that you cannot get. Okay. Including the five that you can see. So it's four nucleotides because there's not a few space. Which is the- Genome iteration
01:20:00
is a post-application modification. It's not a base. It does not exist as a three-nucleotide. You have to have an enzyme that goes into the DNA. Exactly, exactly. You have a slight difference in the way you have it. Yes, because it needs the information on the
01:20:20
other side. It needs the machine like the rest, but also the sandwich movement. It's making transfer rates, yes. So it's going to be very difficult, so different. And these mutations are so congenetic.
01:20:41
Yes, they are, absolutely. So when you get in front of the surprise genetic cycles, it will be an excess of a long, long work of the- Well, it would be tiny. When you get the C to T conditions, absolutely.
01:21:01
But this is completely irritable, and once you have made a new nucleotide.
01:21:27
So when it is starting with the cytosine or the way
01:21:42
for the- Right. That's exactly what, I mean, that's what I did here. I hope, I missed you. And it's a marker that basically what I'm telling you is that genetics is made up of more than just DNA sequence variation. So there is a continuum between epigenetic
01:22:00
variation and DNA sequence variation, okay, to explain genetics, to explain any of it. Yes. Where do you put the partition? I don't know. I mean, if you have been in returns for just three generations, you want to really spend a lot of time talking about this
01:22:21
in a genetic context, where clearly, you know, you assume that things are so stable that you are going to make your cross over a molecular degeneration and still get the same result. Why not? But it is, in many human studies, the only thing that you do is your pedigrees are you, your parents, and your grandparents. Maybe you are a great compound,
01:22:40
and that's about it. And then you go into population studies if you want to go beyond that. So we have this continuum. Well, but, I mean, I'm doing genetics. I didn't show you
01:23:00
any genetic experiment, but really, but that's what I'm doing. I'm doing genetics. I want to understand what underlies the inner returns of character. And we have been told that DNA explains everything. DNA sequence variation explains the differences between characters when these differences are inheritable. And clearly, this is not just that.
01:23:22
Now, the question is, how much of this other system of inheritance which is parallel to DNA sequence variation, how much of this system contributes to the cellular variation that we see? And how much of this system is actually influenced by the embryo? And that's the last question,
01:23:40
which, of course, I didn't, I put that in my summary with the right, you know, the environment. Because to this day, we're completely still, you know, hand-waving. We all assume that epigenetics is so labile that the environment must be what influences it. But the evidence is very poor right now. What we have is, yes, we know that
01:24:00
genes in their activity respond to the environment. You know, heat shock has an impact. You have chemosensitive alleles of genes, et cetera. So we know genes don't function in the assortation. They function in the environment. But how much this environment can now have long-lasting consequences in inducing or erasing
01:24:21
some epigenetic states is still a very, very open question. We tried in the lab very hard to induce the same changes that we have induced through experimental systems just by playing with the environment. And we failed so far. And it makes sense, you know, if we have, you know, this kind of Mendelian genetics at one extreme,
01:24:42
just playing with the immigration, you know, you wouldn't expect that the environment would be able to in between the other side otherwise we would have never had it in the first process. So again, we have a continuous situation. We have a very stable epigenetic state. But exactly what you wanted to do and what you couldn't do can explain specifically. Well, I tried to induce epigenetic variation of the kind that
01:25:01
we have induced using a trick. And I'm not going to go into the trick where I've used it. That's not very important for this talk. So we have tried to induce this epigenetic variation that we can readily implant. Implant, of course. This will implant. Try to do it with this environment. And everything that we have tried to do
01:25:22
failed in terms of transmitting it to the next generation. Well, it failed on two grounds. We failed to induce the same amplitude of methylation changes than the one that we have induced through genetic tricks. So that's the first thing. We haven't managed to find an environment in which we, for instance, can erase
01:25:40
the methylation state of the transposon to such an extent that it doesn't work what we have done using the genetic trick. And second, the few changes that we have managed to induce with the environment in terms of the mutation states over repeat sequences, none of them can be transmitted. Some of them maybe are transmitted to the progeny, to the immediate progeny, as we call parentile effects.
01:26:01
But that's it. Next generation, forget it. It's not there anymore. And that's big. There is a lot of excitement about the environment, but I guess at the end of the day, it also boils down to what people put behind the work of epigenetics. How many people do epigenetics in gene regulation?
01:26:21
And I'm fine by that. Gene regulation, I mean genes, in the expression, are sensitive to the environment. The heat shock response, the gene's antifreeze protein in strawberries, the gene is only when it's minus two degrees,
01:26:40
the gene is off from temperature. So these genes respond through signaling cascades to the environment. The analyzation is another case in France. I mean France are beautiful examples of genes responding to the environment. But all of these responses are plastic and fully reversible. And it makes sense. It's not because you're freezing today that you're going
01:27:00
to freeze in the summer. Therefore, you know, you need now to have your own freeze, but not when the summer is there. At the point of making a protein that costs money and energy when there is no need for it. So far there are indeed just a very short list of non-inherent
01:27:22
epigenetic changes, very few of them. Quite a few. I mean, I mean, I mean, with strong phenotypic effects, probably I'm full of them. I mean, you know, if I were kept so in my reviewing of these cases, maybe we can go to a home right? Maybe?
01:27:40
I don't know. In different species. In different species. But now, with our genome-wide approach in our experimental system, there are thousands of loci potential with equipping direction. And we find that in nature. Now, we don't know how many of these loci do have visible phenotypic consequences because now we have these cognitive traits so it's subtle differences
01:28:01
when they flower, the height of the plant, you know, the weight, the root, the growth. These are very subtle phenotypes. But we can show that. That we say always irrigation, genome-wide, we do what we call this cognitive trait loci mapping, futile mapping, and we can associate differences of motivation with different phenotypes.
01:28:22
I cannot tell you how many genes we will find in ultimately, but the potential is pretty good, pretty large, even for a genome as small as our doses. We saw few repeat sequences to start with. The potential is there, but the inheritance
01:28:41
may be just for a few generations. But, you know, because we are dealing with a continuum, I think that's what is important. It might be for one generation and then forget it. It's not genetic. It's part of LFX. It could be for two or three generations. Ah, it's certainly a bit more complex. It could be 10, 15 generations. It could be 100 generations. It could have mutation consequences
01:29:00
if you have 5 million cytosine. Of course, you will get more often if you are not mutilated, et cetera, et cetera. Okay. So now we have a complexity. We have richness that compose genetic inheritance, which we didn't consider even 10 years ago. And we only consider it now because we have the total sort of consider.
01:29:21
Now we can't sequence the material. We can have, you know, exact mutilation information for individual cytosine. It's complicated actually. It's very simple. So we can, and we have, we have done, we have done tens of metallos, but now a consortium
01:29:40
has done 800 metallos. So you have taken 800 strands of hydrolysis all across the globe, hydrolysis has colonized the environmental industry. So you can fit some strands in China, in Japan, in Central Asia, in Spain, in America, North America. So you have also
01:30:00
coding difference. It's beautiful because in North America, it's only came with the first colon, so 400 years ago. We have a very recent genetic bottleneck and then an expansion there. So we have all of these metallos that are in this cosine and we can see, you know, how stable they are as they fluctuate in relation to climate, environment, but also in relation
01:30:21
to the sequence variation because of course all of these accessions are different genomes and indeed plenty of smiths that distinguish one from the other and things that we cannot explain by either environment or sequence, truly genetic. And we have some because we found them in the lab.
01:30:41
So we go to nature and we find them and now we can start to explore why are they, what brought them about in the first place in nature. And those are stable. Yeah, I'm talking about the ones that are stable and therefore, you know, we want to really pursue them in nature because we know in the lab they are stable and sure enough, the one that are stable in the lab, that's the one that we collect more often
01:31:01
in nature, the one that are unstable in the lab, the one that go back immediately to this default mechanical state while we have difficulties in catching them in nature. And that makes perfect sense if they are perverting, you can't get them in the first place. But you don't understand why some stable and some not. No, at least we don't know and I think we need to do more modeling,
01:31:20
we need physicists or mathematicians to help us with that.