Science & Technology Q&A for Kids (and others) [Part 120]

Hello, everyone. Welcome to another episode of Science and Technology. Q and A for kids and others. I see there are a whole bunch of questions saved up here. Let's see, there's one from John here. Is there a way to digitize DNA sequences and examine them? So we humans, in every cell in our body, we have maybe 100 trillion cells or something. Every cell there's DNA molecule that's all curled up. And the DNA molecule contains a sequence of different units, different collections of atoms that correspond to different so called base pairs. And altogether our DNA is divided into 23 chromosomes. Well, 23 times two chromosomes. But altogether we have about 6 billion of these base pairs. And if we took all of our DNA and just sort of stretched it out, just the single molecule stretched out end to end, it would be about a meter long, but that would be kind of a meter long, but only a few atoms thick, so to speak. So the question is, can we tell what that sequence of base pairs on our DNA actually is? There are four possible bases, usually written as A, T, C and G. And everything about the way that we're constructed is specified by those 6 billion base pairs. That specifies how every protein in our body gets made and how we eventually sort of get put together. So the question is, how do we find that sequence of 6 billion base pairs that represent us? And I don't know, somebody like me, as a practical matter. Well, as a practical matter, it became possible to kind of read that sequence of base pairs in the early 2000s, to sequence this whole human genome. And it became sort of a thing one could do around 2010, I think. And I was in the very first sort of cohort of people who said, okay, let me get my genome sequenced. And it arrived back on a disk drive. Basically, you give a blood sample, it could be a sample of any cells in your body, pretty much any cells that still have their cell nucleus with the DNA and so on in it. So avoid the red blood cells, but any other, like a white blood cell will do fine. Or any other thing that you get from, you know, spitting in a tube or something, you'll have cells that fall off the, the side of your cheek into, into saliva and so on. And that's enough to. You just have to have a cell with its DNA still inside it. And then, then from that DNA, you're finding the sequence and the sequence is then what sort of specifies how, how one works. And one can then go and sort of explore that DNA sequence and find out things about oneself. Like, it was fairly easy to Determine, Yes, I should have brown eyes, which yeah, you could just look in, look visually and you'll see. Yes, that's really true. But that's something that's fairly easy to tell from DNA. More difficult to tell are all sorts of other kind of traits and all sorts of other kind of health related sorts of things. But the question that was asked here is how do you actually find that sequence of base pairs from the DNA? Well, there are a number of technologies that exist for doing that. But let me explain the big point. Big point is you've got all these 6 billion base pairs. There are various technologies where people imagine laying out the whole DNA and just in a very powerful microscope, just looking at what each of base pairs is. That hasn't worked out so far. What instead tends to be done is you take that big sequence of DNA and you break it into pieces. You break it into pieces that may be 30, maybe 100 base pairs long. They're just a lot of fragments. Then it turns out there are technologies that allow you to figure out what the sequence of base pairs on each fragment is. So you have this big bucket of, is it a billion or so examples of this is what this 30 base pair sequence of fragment of DNA. This is its sequence, you know, A, T, G, C, C, G, T, T, A, whatever it is, that's its sequence for all these length 30 fragments. And the way one sets it up is one has chemical and enzyme that basically breaks the DNA up randomly into these fragments. So you've got all these different fragments, you manage to find the sequence for every fragment. You might have a billion different fragments with each with a different sequence. Now the question is, so you've got all these fragments, you know, all these fragments came from somewhere on the original DNA. I should explain another point. When you get the original piece of DNA from that original cell at the beginning, you just have one copy of it. But there's a neat trick called RNA polymerase. There's a neat trick, pcr, polymerase chain reaction. There's a neat trick that allows you to amplify that DNA. And so what started as just one copy, you can end up getting sort of a huge number of copies. And it's from those separate copies that when you smash them into their length 30 sequences, you're smashing different copies into those length 30 sequences. And what you've got is essentially these places where for all practical purposes it was a random length 30 sequence that you snipped out of this length 6 billion altogether. Sequence of DNA. So you've got let's say a billion of these length 30 sequences all floating around. Each of those billion, you know what its sequence was. So now you have to reassemble things because you know that in the end, each of those sequences came from somewhere on the 6 billion base pair, original whole sequence. So it's kind of a puzzle thing that you have to align all those sequences. So, for example, you might find one sequence that has a particular ending to it. You might go attcg, and then you find another sequence which has attcga. I think that's what I said, which happens to show up at the beginning of that sequence. So then you get the idea those two sequences must follow from each other. And so you start being able to assemble the whole DNA sequence from these different pieces. This technique, usually called shotgun sequencing, where you sort of smash the thing apart and then you have to do this reassembly. And one of the bigger problems when this was originally done was that it took a lot of computer effort to figure out what are all the. You might have had to just sort of match every one of those billion fragments with every other one of those billion fragments. And that would be very complicated. But there are better ways to do it so that you can sort of align things so that you can then reconstruct the original sequence from all those fragments of the sequence. So typically, the thing that is the sort of the length 30, let's say block, is you talk about that as a read. There's a single sort of read where you get 30 base pairs read out from that fragment. And then the idea is you try and figure out how would these fragments go together. And you'll end up with a certain depth because you will have said, well, I've got one assembly, one sequence, that one solution to how these things could have gone together here. But another solution, another solution, another solution. So a pretty common thing early on it's gotten bigger now, is it will be 30 deep length 30 reads, which means each sequence is. Each fragment is of length 30. That's the length of the reed. And then the depth is the number of different reconstructions that you've made. Why do you make multiple reconstructions? Well, because all kinds of things go wrong when you do these reads. They're not perfectly accurate. Sometimes an A will be misread as a C or something like this. And so what you then do is you say, well, I've got these 30 possibilities for what occurs at a particular position on my genome. And that. And so what you then do is to say, okay, most of these are T's. All right, that's probably really a T or you know, I can't really tell. It might be a T or it might be an A that's maybe uncertain at that position. And so when you look at a genome that usually there's different kind of letters that are used to encode what possible value there is at that particular position. So it might just be, it's a T or it might say, I think it's N, which means one has no idea what that particular, what that base pair at that position is. Or it might be something else, which means it's either a T or an A. And so that's the thing you're trying to put together is you make what is called the consensus sequence, which is the one which comes from sort of typically the largest number of possible. Largest number of. If it said it was a table, but that was the most likely thing it said it was, then the consensus sequence will be one where you have say it's a T at that position. Now it's kind of tricky when you specify a genome sequence because everybody has, okay, so there is a sort of standard sequence which is kind of, well, there isn't a standard human sequence because we're all different. But there is a, you might imagine, isn't really true, that there's sort of an average human genome sequence. Not really clear what that means because you know, there's 8 billion humans right now and you know what's really the average? There's really in a good average. When this was originally done, the main person who invented shotgun sequencing actually a large part of his DNA was the standard human sequence, so to speak. But if you say for each given person they differ from the standard sequence, the so called reference sequence, at any given time there's usually a reference human genome. I think we're on if it's still build 37 of the reference human genome. I'm not sure. Every so often there's sort of an update to what the reference human genome is. But what you do when you talk about your particular genome is you'll say, well for example, how does my genome differ from the reference human genome? Most common difference is a single nucleotide polymorphism, a snp, an SNP which is a place where there was a T in the reference sequence but you had a G there, for example. And so you'll note that there was a particular snp, a particular change. And we all have, what is it, some number of millions, tens of millions of SNPs relative to the reference human genome. We all differ from the reference by lots of single point changes. There are not just single point changes. There are also sometimes whole sections of the DNA that are missing in a particular person or additions to a particular piece of DNA. Or sometimes there are pieces of DNA where there'll be multiple repeats in one person and only one repeats in other people. One feature of these cases where things are missing or where things repeat or where there's an extra base pair added, is it means that if you try and line up sequences, if you try and line up your sequence to the reference sequence, you don't get to say at position such and such on chromosome, the DNA is broken up into these. A particular person's DNA is broken up into these 23 chromosomes, but that's just pieces of DNA that are separately exist in cells. But on chromosome 1, we might say at position 5 million 742 or something. What do you have there? Is it a T or a C? Well, that isn't completely meaningful because any given person might have had additions, deletions and so on ahead of that 5 millionth position, so their 5 million won't align with the 5 million of the reference sequence. So it's a tricky business to try and say, you know what to try and label? Oh, it's this SNP that is @ this quote's position. And that means that every SNP typically gets an ID number, It's usually RS and then some series of digits that represent that particular snp. And one knows on the reference genome, one knows where that SNP is in your particular genome. It might have moved around relative to where it was in terms of coordinate position relative to the reference human genome. And typically the way one deals with that, I don't know whenever I go look at my genome, which is fun to do from time to time, when you see some grand new piece of research that says if you have this snp, then it means you're going to have this trait or whatever else. How does one actually find something in one's genome? I use, in Wolfram language, we have a nice way of representing genomes. And essentially what I do is I always say, well, let's look at the reference human genome. Let's look at the flanking sequence around that particular position. So let's say around that particular position, things always go atcg, gtca, whatever on one side and then some other sequence on the other side. Well, as soon as I've got enough base pairs that I know go on either side, I can basically Just search for those flanking sequences in my whole genome. And the chances are the only place where those flanking sequences will occur is right around this position that I'm looking for, where the SNP occurs. So that's a convenient way to kind of align oneself with a particular position on the genome. And by the way, that kind of mechanism of looking for sequences like that, the immune system makes use of that. Lots of things make use of this idea of. In the immune system, one breaks pieces of viral genome, for example, into, I think they're length 12 sequences, something like that. And one's using the fact that if you've got this length 12 sequence, you can kind of tell is it part of me or is it not part of me? Because In a length 12 sequence, there are four to the 12 possible sequences made of base pairs that are length 12. And you can kind of say, well, of those, only a very small number will be actually present in my 6 billion base pairs in my genome. And so you can use that, and the immune system uses that in a very fundamental way as its way of detecting is it something that came from your genome that's part of you, or is it some alien thing that came from a virus or a bacterium or something else that's an alien thing? Not that has a sequence, that fragment of the sequence has never been seen before. So in any case, sort of the zooming out a bit, the main thing one's doing is you've got this DNA. You want to figure out what the base pair sequence is. You break the DNA into these little sections, you, you read each section, then you align the results you get from all those sections to be able to make this whole long, continuous 6 billion base pair thing. And then I'm sort of describing how you make use of what you get from that. I will say that the technologies for reading, for making those length 30 reads, there are several technologies there. And some of them involve, they basically involve sequentially going down those little fragments of DNA and sort of watching one base pair at a time what happens in some chemical reaction. I think there's. The main technology, I think uses glass beads with different kinds of materials on those glass beads that basically, as you're sort of walking down this piece of DNA, you see different signals depending on whether the thing that came next was a C or a G or whatever else. There's a nanopore mechanism where you are kind of pulling the DNA through a membrane and you're sort of watching what happens as each different base pair kind of goes through the membrane. There were some very direct techniques that people tried to develop. One of the ones that was quite fun was to literally lay the DNA out in some kind of plastic container and literally look at this DNA that's been laid out through an electron microscope to see what the base pair sequence was actually like. That technology never came to fruition, at least not yet. But that's a type of thing one can imagine. And in general, a sequencing machine. I don't think I've seen a sequencing machine in a few years, but last time I saw them, they were, you know, in a bio lab. There's a certain tendency for everything to look like a kitchen appliance. And, you know, sequencing machines kind of like you put the DNA in and it's this thing, it's this box. And there are different technologies that allow you to sequence more quickly, but not quite as accurately or sequence accurately. And it takes hours or even days, I think, to get sequences out. And then the amount of data you get from these sequences is quite long. If you've just got the consensus sequence, then that's just 6 billion base pairs. But if you've also got every different read, then that's much more data. I think nowadays, I think I just had a bunch of. Because I like these things, I get my genome sequenced every so often with ever better technology. And I think nowadays it is length 100 reads and I think they're like 100 deep read. So there's 100 copies of each of each thing that's read. I think I will say that there are places where all kinds of things, there are all kinds of funky issues. Like for example, whenever there's repeats in the DNA sequence, those are hard to deal with with this shotgun sequencing technique, because once you have enough repeats, if it just goes 80, 80, 80, 80, 80,. Et cetera, et cetera, et cetera, you kind of don't know how to align things. There's no way to know when you've got all those at, AT, et cetera, together. It's like, how long was that piece and how did this at, at @ align with that same repetition in some other fragment? And, and so that causes trouble. And sometimes when there are genes, genes are the, the pieces of DNA that will make single proteins. So there'll be some run of, of parts of the, of your genome that are, are what are used to make a particular protein. So in the end it's transcribed onto rna. Then the ribosome that kind of walks down the RNA strand, extruding pieces of protein. It knows to kind of start reading at a particular place on the DNA. There's a so called start codon there. And it keeps going until it gets to the stop codon and it stops. And that's the end of, that's the end of producing that particular protein. But sometimes there are genes where in the history of biological evolution, those genes were very similar. And so while they might have ended up in different places on your genome, they're very similar. And that causes all kinds of confusion for this shotgun sequencing technique because it's like there are two parts that really look the same, but they're actually from different pieces of the genome because those two genes had some kind of biological evolution convergence, so to speak. Another thing that happens is the ends of DNA are these things called telomeres. And telomeres are repeating blocks. Usually if you're kind of, if your genome is in great shape, you'll have about 50 telomeres at each end. And they kind of hold the end caps of the DNA. As DNA replicates or you get older, presumably there are fewer telomeres. And although that's been sort of an issue of, you know, do you just lose telomeres as you, as there are more replications of your DNA and as you get older and that's so. And then your DNA sort of unspools and that's bad news. Or can you eat well, you know, be healthy? And actually there's a thing called telomerase, an enzyme which adds telomeres. And can you actually just, you know, keep the telomeres at reasonable length? But in any case, one feature of these telomeres is that because they're repeating blocks, they really can't be very well read by this kind of shotgun technique. And so sort of, if you want to know how many telomeres do you have left at each end of your genome? That's sort of a separate thing that you have to work out differently from the main sequencing process. So anyway, that was a little bit on how DNA gets some. How to sequence DNA, how to find the sequence of base pairs corresponding to your particular genome sequence. Let's see. Okay, let's see. There's a question here from Philip about is a complete family tree of humanity a realistic possibility? So, you know, there's 8 billion humans right now, give or take, about 100 billion humans probably have lived in the history of our species. I think that to reconstruct what the family tree of all of those humans is seems like a realistic possibility. There are some places, like I think famously Iceland, which was Settled in, What, a thousand A.D. or something, where pretty much the complete genealogical history of the population there is known. And that was done at some point. I don't know. When was it? It was about 15 years ago, maybe. There was a big effort to. You don't need to sequence everybody. If you know enough about kind of the birth records and so on. You can just sequence particular people and then you can kind of deduce what the intermediate people must have been. And occasionally you find all sorts of strange things where this sequence couldn't possibly go with this person being a sibling of this person. So something was wrong with the birth records and so on. But I think that it becomes perfectly possible and I think it will be, you know, it will gradually be doable to really fill in the sort of pretty much complete genealogical tree of our species. Now, some aspects of that are pretty well known already. There are. You can deduce things about sort of where did your ancestors come from in the world. And there are so called haplogroups, which are kind of a way to characterize sequences in the DNA to say, oh, these folks came from, you know, they were indigenous people somewhere in the Americas or these people were from, you know, southern Europe or whatever, some amount of time in the past. Why does that work? Well, so, okay, so usually there are two pieces that one looks at. So the mitochondrion, which is the energy source for cells, but it's a piece of our cells that replicates independently of the main cell and it has its own DNA. So mitochondrial DNA is passed down the maternal line. It's passed from mother to child and so on. And so for us males of the species, our mitochondria are kind of, you know, it's the end of the line, so to speak. We don't pass our mitochondria to our progeny. But so if you follow the mitochondria, you're following the female line of a particular genealogical line, so to speak. And then for males of the species, the Y chromosome in the X and Y chromosomes. Females of the species have two X chromosomes. Males of the species have an X and a Y chromosome. That's a particular human feature. If you look at other organisms, you'll find different arrangements of chromosomes and different things that lead to kind of the different male, female, et cetera, configurations for different species. But that's what we humans have. And so if you look at Y chromosomes, those are passed from, you know, from father to son, so to speak. And that's a different line. But so when you look at mitochondrial DNA, for example, you're able to sort of see back down the female line, so to speak, what were the origins? Well, what is, why does that work? It works because when you where each generation, there are small changes in the DNA, more or less, you get sort of half your DNA is from your parent, from one parent, half your DNA is from your other parent. And if you look at a particular chromosome in the DNA, what you'll see is that there'll be a run of stuff that comes from one parent and then a run of stuff that comes from another parent and so on. And I think they're usually about, so, I don't know, half a dozen, maybe a dozen crossovers on a single chromosome. So you'll have a big run of, you know, sort of father produced DNA, big run of mother produced DNA and so on. That means there are sort of traits that tend to be in common from one parent. But it's very hard to see that because the, on, on the, it's kind of like the functionality of the DNA, what genes produce, what proteins and so on, it's all been kind of jumbled up in terms of the sequence. So the, you know, the gene that encodes things to do with, I don't know, eye color, we write next to the gene that encodes something to do with, you know, cells in your kidney or something like this. And there's no, it's not like, oh, this section, it's not a well organized thing where it's like, oh, this section involves the brain, this section involves the foot, things like this, it's all jumbled up. But in any case, the point is that as you go from half your DNA is from one parent, half your DNA is from another parent. But there's also, every time DNA replicates, there's a certain rate of making mistakes, there's a certain rate of where it should have been a T, it ends up being a C, for example. And we all have about, I think it's about 700,000 SNPs, 700,000 places where one mistake was made relative to either of our parents, so to speak. So we have about that number of sort of never seen before changes to our DNA. And those never seen before changes are important. They might be really good, they might be what makes you unique and great or whatever, or they might be something that causes trouble. But in any case, they're things which are changes. And if you look over the course of many generations, you'll see those changes progressively building up. So you'll go from Something where all the DNA is just like it was for this ancestor, and then it progressively changes. And that allows you to kind of figure out sort of things like how many generations was it from this original DNA back here to the DNA that we now have today. But it also means that because those changes, 700,000 is pretty small compared to 6 billion or 3 billion from each parent. It means that you can pretty well trace what the parentage was, even independent of those changes that are getting made. And so that means that you can trace back where you can see sort of what was the DNA of the ancestors. And then what people have done is to try and compare the four particular DNA. They ask sort of where were the folks who had that DNA, where did they live in the world? And so you get these maps of where you know, your DNA, where did it originate in the world? And you can also see in the course of sort of the evolution of the species who mated with who to mix in DNA from this place and that place and so on. I mean, most of human history, most kind of mating was pretty local. I mean, it's only in fairly recent times that it's like there'll be children whose parents come from vastly different parts of the world. And so then that allows one to deduce, oh, you know, with my DNA, I can tell my ancestors came from this particular part of the world. And there's been quite a bit of work that's been done on tracking sort of migrations of large populations back in the past. And you can say, well, you know, 20,000 years ago, my ancestors hung out in this part of the world, et cetera, et cetera, et cetera. And that's something that is, at this point, a pretty well developed technology. Now, occasionally it'll be like, well, you're 99% come from this part of the world, but there's 1% that's some utterly exotic thing from some completely different corner of the world. That's usually a mistake. That's usually because there was some change in the DNA. And the algorithm that was used to decide which of the different possible places where your DNA might come from was it from, didn't take account of that sort of unusual change. And so it said, well, you're 1%. Some completely weird place that you can't imagine any of your ancestors ever had anything to do with. So let's see. So, I mean, that's one part of the kind of genealogy of our species that's at this point fairly well known as these, these kind of large scale Kind of groupings in these haplotypes and so on. And then in terms of really being able to fill in the finer details of sort of the complete net of the genealogy of our species, I'm sure that will become possible and I think it'll be kind of amusing. I mean there's a lot of, well, for example, there are particular, I don't know, Genghis Khan for example, from what the 1200s, 1300s had an awful lot of children. And so there's an awful lot of people who have some sort of piece of Genghis Khan derived DNA. I think somewhere in Ireland, I think there's another one of these very prolific sources of DNA. But there are things like this. But I think it will be possible undoubtedly to sort of reproduce the whole net of connections there. Let's see, Let's see. So one, there's a question here from Prab. Is DNA a tree or a semi lattice? Wow. So you know, when there's a piece of DNA and you're passing it from parent to child, the DNA that the child gets is slightly, slightly different from the parent, as I mentioned. Now it's always possible that sort of somewhere down the line, you know this, the, the children of the children of the children on this side will end up having the exact same DNA as the children of the children and children on the side. But it's unbelievably unlikely. I think there's a fun movie that has some science fiction kind of concept that there'll be kind of a galactic monarchy or something where the next monarch is chosen based on them happening to have the exact same DNA as the previous one, so to speak. I think one will be waiting a very long time for the next monarch in that scenario, if one was doing it with the full stretch of DNA. So the chance that there's sort of a convergence is sort of a chance. Convergence of a large chunk of DNA is really low. The chance that. So what tends to Happen is these SNPs that I've mentioned for various reasons, which I'm not sure whether they're completely known. There are some places that are kind of more likely to have errors of transcription, errors of transmission to the next generation than others. There are places that are more likely to have sort of changes than others. And there's some evidence that there are in particularly in lower organisms, maybe viruses, things like that, that there are sort of. It's in the interest of the organism to evolve, to change, to adapt to the environment and sometimes environmental factors and sort of stress from the environment changing will cause the DNA to tend to evolve more quickly, for there to be more errors of transcription and so on, because it's sort of in the interest of the organism to seek out different forms, so to speak. Because, you know, the environment has changed and you've got to adapt. And if you just stay the same as you were before, that's not a good strategy. So I think. But the idea that there would be sort of major convergence of different DNA from different branches of the tree is very unlikely. Now having said that, when it comes to, if you look at the tree of life and you ask, are there mechanisms first thing to realize about the tree of life, the sequence of organisms going back whatever 3 billion years or something to the very first living organisms on Earth. One of the things that's always utterly shocking about the tree of life and about the things that we have is we say, well, we've got this mechanism, we've got this thing about how brains work and it's this particular protein that used in brains or it's this particular mechanism of how sleep works or whatever else it is. And then you ask yourself, how long has it been like that? And you realize, well, actually it's been like that for 500 million years. The first time that particular mechanism existed was really long ago. And we might say we are sort of these wonderful, I don't know what a few hundred thousand year old species. And it's, you know, we should be very impressed with all the attributes we have, but an awful lot of the mechanisms that operate biologically for us are extremely old in terms of biological history. Now, you know, one of the things that happens is there are things that get invented like, I don't know, eyes get invented, things that detect light, other things, you know, walking gets invented. Well, some of those things, they've been invented more than once in the tree of life. Like eyes, I think, have been invented maybe four times and you can see that they've been independently invented because the detailed engineering solutions that have been adopted by different organisms are completely different. And you wouldn't expect. And you can't find sort of a fossil record that takes you from here to there. So there is this phenomenon of convergent evolution where some mechanism like eyes gets to arise because, well, it was a good thing to do if you were a critter living on the ocean floor. It was a good thing to do if you were a flying critter. You needed eyes in both cases, but they weren't sort of connected in the tree of life. One didn't evolve from the other. They independently discovered the same mechanism. So that can happen. But if you ask what are the genetic aspects of how those things were constructed, they'll be completely different. It's like two programs that do. The two computer programs that do pretty much the same thing, but they were written by different programmers and they work in completely different ways now. Yeah. So, I mean, there are mechanisms that will be in common because they're sort of lower level mechanisms. Like there's a mechanism when a biological organism grows. There's a mechanism by which different types of tissue can be, you can differentiate into different types of tissue. You can end up with something as bone, something as muscle, whatever else. And that differentiation process is governed by an incredibly ancient mechanism that's existed for, I don't know what, at least a billion years, more than that, probably. And that mechanism will be in common between the critter that discovered eyes that work this way and the critter that discovered eyes that work that way. Let's see, there's a question here. How likely is it that genetic engineering can create mammalian species with superhuman intelligence? So one question is, if we look at, well, if we look at biological evolution, what mostly is happening is the critters that are more successful in the environment tend to have more kids. And so whatever traits they have, they get passed to more organisms and their kids, and then that keeps going. And it's a pretty surprisingly quick process, sort of mathematically, that if there's a small advantage, you know, if this trait causes you to have a few more kids than this other trait, the trait that causes you to have more kids will win out in the population because there'll just be more and more and more organisms that have passed that genetics down to their progeny. And so that will tend to take over. But that process kind of only works. Well, it depends on what you mean by the environment, whatever the environment is, if it turns out that folks who get more education have less kids or something, then that will cause it to be the case that that will be disfavored in the population in the end. And in modern times, some of the traits that would be like, oh, if you have that medical condition, you're not going to make it to the point where you have kids, for example. Well, with modern medicine you will make it. And so that trait that would have been something that would have died out in the population, no longer dies out because fortunately, medicine has allowed people to live longer with that trait and potentially pass that trait to their kids. So the thing so then a question might be, well, what are some traits that might arise in the population that would be sort of more successful and that would cause sort of, you know, traits where people will have more kids if they have this trait? And that's an interesting question. Whether there is evolution still going on in the human population, there presumably still is. It may be rather bizarre because unlike other times in biological history when the evolution was, can you get enough food and can you pick the right nut from the tree? And whatever else it is, the forces of evolution in today's world are very different and often have economic characters and other kinds of things that are very different from the ones that have existed kind of in the wild, so to speak, in the past, although there are presumably still forces. And one can ask, what are traits that you can see in the population today. That seem to be better than a small percentage trait that seems to be better than other people have? There's one that I know that has to do with red blood cells, carriers of oxygen in the blood. And it's a trait that causes some people to have a higher concentration of red blood cells than most of the population. And actually, this trait was discovered because one of the things that if you're an athlete and you want to have more endurance and so on, you want more red blood cells. And you can get that by. By basically injecting erythropoietin epo. It's a substance that causes the production of red blood cells. And people have kind of outlawed doping with that substance as a way to kind of improve your athletic performance, because that's thought to be not a fair way to be a human competitor, so to speak. But there was a particular individual who did really well in a bunch of athletic competitions that people were saying, you must be doping with the CPO thing. And it turned out, no, he wasn't actually. He just had a genetic modification that caused him to produce more red blood cells than the typical person. And presumably having more red blood cells is an advantage. There doesn't seem to be an obvious downside to it. It just seems like it gives you more endurance and so on. So that's an example of a kind of a trait in the population that I don't think it necessarily leads to more kids, and therefore it's not going to by natural selection. It's not going to sort of take over the population. But it's an interesting trait, if you have it, so to speak. Well, the thing that. So the question then is, well, all right, let's say there Are traits where you say, this is a good trait. And now you start saying, well, should we sort of engineer that trait into kids of the next generation? Very tricky issue. I mean, that's a very tricky, messy story because there are cases where you say, well, gosh, if there's some nasty genetic disease that's produced by some particular feature of DNA and you detect that at the level of. You detect that in, for example, what would become a fertilized egg that's going to turn into a fetus and so on, if you can detect that and you can go and make modifications to the DNA sequence and you can remove that genetic defect, surely you say that's a good thing. Presumably, yes, because it will prevent that. Otherwise that child would have this nasty genetic disease, might not even, you know, survive to birth, might have all kinds of other problems. And gosh, we can fix that. We can just make do. Gene editing. I should explain that something was discovered about. When was it, 10 years ago? A little bit more than that was the ability to essentially just edit a piece of DNA to go in to your DNA. I can explain some other time how that works, but basically what's happening is you send in this kind of molecule that essentially searches out these flanking sequences. I was talking about these flanking sequences away. If one's trying to find a particular SNP in one's genome on a computer, you look for, search for atgc, whatever else it is. Well, gene editing does the same thing. You give it these, I guess they're primers, these sequences of little fragments of DNA sequences. And you say, okay, when you see these two flanking sequences go in and edit the stuff in between them and change it from this to that. It's a very kind of. It's kind of almost like a search replace, text editing kind of operation. It's very similar to what you would see in like, well, pattern matching in Wolfram language. A bunch of other different places where you see that same kind of, you know, you see something which looks like this, make a change to, to something associated with that. Anyway, with gene editing you can potentially. Make changes to DNA. You can edit things. And this was done once and everybody got very upset about it. And it's not clear what, it's not clear what should be done, how it should be done and so on. One of the things that tends to happen is somebody says, oh, it's terrific. Let's just make this change to DNA. Let's change it so that everybody has blue eyes. Let's say, well, might turn out that there's some unintended consequence that if you do that, then that change in the genome also has the effect that, I don't know, people will tend to develop some particular disease or some such other thing. It's hard to know right now what we're doing is we're leaving it to fate and nature and so on, kind of rolling the dice and saying whatever nature does, we'll be okay with that. What one is asking about is whether if we know, sort of engineering wise, if we say, well, we think based on this engineering calculation effectively that this will produce a better result, we can engineer the better human, so to speak, should we do that? And what is if we think we know what's going on and can say this will be the outcome, is it worse to do it when we think we know what's going on than when we're just rolling the dice? Unclear, complicated story. But in terms of the traits that lead to intelligence, whatever that means, I mean, that's a particularly complicated mess because there are, there are surely, you know, first of all, it's very unclear what we mean by quotes, intelligence. I mean, there are, there are, you know, we have, you know, one of the things we learn from computers is given any computer, you can program it to do anything that any other computer can do. And I'm pretty sure it's the same thing with brains. Given any brain, you can program it effectively to do what any other brain can do. It might not do it quite as quickly, it might have certain other attributes, but it's pretty much sort of universal, so to speak. But then the question would be, well, okay, which brain is faster at doing which thing? Which brain has this or that trait? And surely there are differences there, just as there are differences in the heights of different people and other characteristics. But saying generally, people had this idea back 100 years ago or so. Now, this notion of the quotient of general intelligence, so to speak, it's not clear how meaningful that is, because it's not clear when you say, well, I want to use my brain to do this particular thing. Well, you might be really great at doing that particular thing, but really crummy at doing this other thing. What does that mean for your general intelligence quotient? I don't think the general really survives. It becomes, are you good at doing these particular kinds of visual IQ tests or something? And that is, well, you can measure that. And some people be better at those than others, just as some people run faster than others. But that doesn't to then conclude that there's A whole. Well, then the real question is, if you look at different occupations and different things people do in the world, how aligned are they with different kinds of attributes that you can measure with different kinds of tests and so on? And the answer is, it's really complicated. I mean, when you look at, oh, I don't know, let's say, doing mathematics at a research type level, one of the things you notice there is you say, well, if you're going to do mathematics at a research level, you have to be good at blah, blah, blah. You might say you have to be good at dealing with numbers or something. Turns out it isn't true. It turns out there are a whole bunch of very different ways to do sophisticated research. Mathematics. You can be very good at thinking geometrically. You can be very good at thinking kind of symbolically. You can be good at seeing patterns between this and that. There are different ways to do it. The end result is you do good math. But how you get there and the kinds of explanations that you give to yourself and other people may be utterly different. So to say, well, we measure this particular one trait, and that's the thing that determines are you going to be good at math? It's not going to work that way. It doesn't work that way. So if you say, are there changes you can make in biological changes that can lead to changes in overall intelligence? It's unclear what that means. But there are surely things that in principle you could measure them. We don't know what pretty much any of those traits are genetically, where you could measure it and say, okay, if you have more of this, if your gene for this has this characteristic, then you're more likely to be able to do this kind of puzzle quickly. Now, having said that, things like that tend to be really hard to figure out from genetics. There are some traits like eye color, that are pretty straightforward, where you can just look at individual little regions or individual SNPs on the genome and you can determine it. But, for example, height is one that people have looked at extensively. And it isn't something where you just say there's a gene for height. If you have this variant with ATC here, then you'll be at least 6ft tall. And if you have this variant, then you'll be less than five feet tall. You know, there isn't any known such thing. And presumably what happens is that there are a whole collection of different genes, hundreds, maybe thousands of genes that all contribute in some way or another to overall height. There's some evidence for that because if you look at the distribution of human heights, it follows a normal distribution. And that's a typical thing that you get when you have many independent variables that are all being averaged together to get to the final result. It's like there are many independent variables that determine in one way, another, oh, your leg bone in this way is going to be slightly longer. Oh, your spine is going to be slightly longer or shorter in this way, et cetera, et cetera, et cetera. And that that will be. It's a combination of all those things that leads to height. But so I think I have to say, if I were going to sort of say what's likely to happen in practice in the kind of genetic, you know, genetics of humans and intelligence and so on, I have to say, tool use is a much better piece of leverage for improving the effective intelligence of humans, I suspect, than going in and doing potentially very crazy things with the genome. I mean, it's much easier to say, gosh, you know, this type of thing that used to require some particular intellectual trait to be able to do it, actually, now we can just get a computer to do that. We don't have to know how to do it. In my own life, you know, I got interested in doing physics when I was a kid, and I didn't like doing some of the sort of mechanical aspects of math that were involved. And so I started using computers to do that. And then I built a bunch of tools which fortunately, lots and lots of other people have now had a chance to use, too, to do mathematical calculation with computers. And now at this point, it's like, back in the day, if I hadn't been able to do certain kinds of algebraic computations, I would have been sort of locked out of doing various kinds of physics. But as it was, I got computers to do those things, and in fact, they did them a lot better than people could do them. And so I was not locked out. I was kind of pulled in to doing a lot of things that, that weren't really easy for other people to do because I was using computers. So it's kind of like instead of trying to improve yourself, improve the tools you use, use the tools. And that's sort of the faster way ahead, so to speak, than, oh, let's change the genome and see what happens, so to speak. Plus, you don't have to grow from, you know, you don't have to biologically grow to make use of tools. You just have to choose to do it. I suppose that would be a pitch for my life work of building Our computational language as sort of a general tool to let people do with computers and sort of extend the things that they think about by using computers. And I think that's a more sure path to sort of effective enhanced intelligence for humans than anything that has to do with. With messing with the genome. I mean, the history of trying to do things like that is fairly terrible. I mean, at various times, people have said, oh, there's this collection of people that have this great trait that have whatever this is in history. This has tended to end very badly with a lot of bad things happening. And at this point, I think not a lot of enthusiasm for kind of this kind of genetic selection or genetic editing or whatever else. But it's a tricky thing because in a case where you can say, this is a person who's going to die at a young age unless you make this genetic change, it's kind of like you might say, well, you should make that genetic change. And there's a good argument for that. But I think there's perhaps a. It becomes more complicated when that change will get propagated to future generations, which it will typically. So this is a tricky business. Let's see, There's a question from Zayden about epigenetics. Oh, let's see. I want to get into that. Boy, we've really had a genomics day today. Okay, let's talk a little bit about that. So one of the issues is you have a genome, and that specifies kind of what proteins you'll make, which specifies sort of what you as an organism, how you as an organism are constructed. But there are other factors. There are other things about, for example, your environment that can affect aspects of how sort of how much of a particular protein is made, whether that protein is made, whether there are modifications of that protein afterwards, whether there are even modifications to your DNA, not to the base pair sequence. But there's this process of methylation, which is a way of sort of adding annotations to the DNA. And that seems to be. Methylation seems to be quite important. It seems to be important both in the ordinary operation of organisms and to things like cancer. And that's a place where even though you've got this, your standard base pair sequence, exactly what that methylation looks like can affect how things work. And so there's a whole sort of separate world of, okay, what kind of sort of after the pure genome, what kind of modifications get made? I'm not sure I can quite tell you in an organized way. There are lots of different things that are known about how that works. One of the key problems is everybody starts from one fertilized egg cell. It splits in two, splits in two again, splits in two again. Eventually you get the hundreds of trillions of cells that make up each of us. One thing that happens is those cells form different types of tissues. They might form a few hundred different types of tissues, Whether they're hair cells or skin cells or, or lung cells or brain cells or whatever else, they differentiate into different kinds of tissues. And there's a long time question of how does that differentiation work? How is it the case that you can go from sort of the original stem cell, the original sort of source cell, and it differentiates into this kind, that kind, that kind. And there are, if you look at that sort of tree of differentiation, I think for us humans, it goes about 3, 15 levels deep. I think that, you know, if you're trying to produce a cell of a particular kind, there's sort of 15 choices that you have to make to get to that particular cell, as opposed to becoming a heart cell or a brain cell or whatever else. It used to be thought that it was very difficult to go back to a so called pluripotent stem cell, a stem cell that has many possibilities for what it will turn into. It used to be thought the only way you could get those was, was to go back to fetal tissue. But turned out that it was figured out how to kind of take any cell and put it back into the stem cell state where it's capable of sort of going anywhere. And the question is, well, when it goes, when it becomes a particular type of cell, what does it mean to become a particular type of cell? What it means is in a particular type of cell, the proteins that are relevant for the function of that cell are being produced a lot, and the proteins that are relevant for the function of another cell are not being produced, produced. And I think, I'm not sure what the latest story on this is. I think there's a bunch of methylation related stuff that's known about sort of how that differentiation process happens. It's been figured out in fairly recent years, but that's sort of one example of not just the genome, but also it's your same genome, but it's a heart cell. And that's an example of one of these things that's sort of not just genomics. And so, for example, there are things that can affect, for example, if you have identical twins, they don't necessarily have everything the same, because if it was just a question of their genomes their genomes are the same, but there are aspects of identical twins that aren't exactly the same. Like for example, even things like their fingerprints may have maybe similar, but probably not exactly the same. The shape of their ears, the kind of the curling up of the layers in their brains and so on may be similar, but not the same. And part of the reason for that is that those things like fingerprints and so on are the results of the particular mechanical folding of tissue when things were growing. And that particular folding of tissue was not determined by the genome. Maybe some global aspects, some coarse aspects of it were determined by the genome, but the details were not. And so that's something where even with identical twins, they won't end up being exactly the same. And that's what's sometimes called sort of epigenetic phenomena. And there are different levels of things that actually go down to affect things that are sort of direct modifications of the genome versus things that are much more mechanical, like did it fold this way or that way? So a few thoughts about that. Oh boy. All right, a few more questions. Arc asks, do viruses play an important part in evolution? I'm not sure we completely know. I mean there are so called retroviruses that go and insert themselves into one's DNA. There are viruses which take a lump of virus DNA and just, they are the analog of what this genetic editing, gene editing process is. They just go and insert themselves into your DNA. And I kind of wondered because I've got my genome sequenced first in 2010, I think, and now we're 13 years later and it's like if you look at my genome today and my genome back then, if I got sick with a virus and it happened to be a retrovirus, would there be actual changes in my DNA? Well, the answer is the, not most of it because the virus didn't get into most of the cells in my body and most of the cells it did get into my immune system will have killed off. So that probably isn't a story there. Probably. But is that something that's important in the overall history of life on earth? Undoubtedly, yes. Undoubtedly it's important because among other things, viruses have much more rapid DNA, much more rapid genomic change than typical organisms. They have much shorter kind of replication times, they have much less DNA. So they don't have, doesn't take as long to replicate that. And so they're probably a fast running place where changes can be made. And if they can insert themselves or even possibly their interaction with defining the environment, the fitness, I mean, it could be the case. Some virus comes in and it evolved in a particular way, and it wipes out some part of the population that would otherwise have been considered quite fit and quite able to produce lots of descendants. And so that's another mechanism by which one can expect that sort of the evolution of viruses is correlated with the evolution of other organisms. Boy Gal asks, why is it that mammals like whales and dolphins made their way back to the ocean to evolve as creatures living in the sea rather than staying on land? I don't know that history in detail. I know that evolution is a very gradual process, and when things have to happen suddenly, evolution isn't really able to do them. So, for example, if you look at the first fish, and when was it? Maybe Carboniferous period or something? When. No, it's earlier than that. I think when there were first sort of creatures on the land, it's the fins were like, okay, you can kind of push yourself on the beach with your fins, and then, oh, no, now they're legs and now you can walk on the land. And I think if you're a, you know, if you look at the structure of the fins in, in something like, you know, whale, dolphin, whatever else, their bone structure is very similar to the bone structure of our hands and feet. They just have, you know, they're bigger. And so at least whales are. And so they have more, just as we have. How many segments is it like three segments in our fingers? Their, their finger analogs have like, I don't know, 15 segments or something, but they still have the same basic bone structure. And it's, again, it's this terrible feature or good feature. I don't know how you think about it, of biological evolution that there are very few new ideas. So it's like we've got the idea of a hand. We've got this idea of a thing that is sort of a flat, modifiable thing. Okay, let's use it as a flipper. Let's use it as a hand. And for us, when our hands grow, they actually start with webbing between the fingers, and the cells that form that webbing die off before we're born. And that's how we end up with these separated fingers. But kind of the idea of, can you make a web thing that has the same bone structure? Well, yes, you can. Now, why those organisms sort of went back in the ocean from the land, I don't know. Usually these things, you can kind of tell a story in retrospect for why something happened, but it may not be quotes, really. Why it happened. I mean, for all one knows, things got cold on the land and so then it was warmer in the ocean. I don't know. Or it became worthwhile to be bigger and bigger and bigger. And by the time you were that big, it was hard to walk around on the land. All sorts of different things like that. I don't know. And there may be no good answer other than an answer in retrospect where you're kind of telling a story that's just about what actually happened and you're not really explaining quotes why that happened. I don't know what this means. The. Okay, maybe one last question here from William. Another biology question, but a little bit different than we've been talking about. How does the brain distinguish signals coming from different senses? What's the difference between signals coming from the eye versus the ear? Is everything ultimately an electrical signal? Is it not a difference in. It's a difference in degree instead of in kind. You know, in a sense, the answer that is very, very simple. You know, we have a bundle of nerves that go our optic nerve that goes from our eyes to the visual cortex to the back of our brains. We have a bundle of nerves that come from our cochlear and our ears to the auditory center in our brains. And in a sense, they're just different wires. We know that it's vision because it's coming down the vision wire, so to speak, to that part of the brain. We know it's sound that because it's coming down those wires, in effect, the signals are very similar. I mean, I think in the, you know, you might have thought, oh, sound will be. It'll record just like on a computer. It'll have direct data on, you know, the eardrum was pushed in, out in, out at the speed. No, that's not actually how things are encoded. What's done is the frequencies are separated by, you know, hair cells at different places in the cochlea move at different. When the sound has different frequency components and by the time it's being transmitted to the brain, it's really, oh, in this sound there was this intensity at this frequency. It's very, very similar to the way when we're looking in vision we're saying, oh, this photocell. Photoreceptor on our retina was excited by these photons coming from this angle. It's. There's something going on to the left. There's is very similar to. There's something going on at a high frequency there. But in the end, all of these signals are Just electrical signals in nerve cells in these chunks of nerves that come from our different sensory organs. And I don't think there's a significant difference in what you see. I mean, in other words, a nerve cell that's been excited, it's kind of like producing electrical signals, usually about a thousandth of a second apartment. And if it hasn't been excited, it's just hanging out and occasionally making a little bit just for fun. And no doubt it isn't just for fun. No doubt it has some biological origin or some electrochemical origin. But in any case, the thing that. The way we know what scent it's coming from is literally what wire it came down. And the same is true for touch. For example, we're literally mapping, you know, when. When you touch your finger, there's particular nerve endings in your finger, and there's a particular nerve fiber that will go. And it will be the. It won't be a single fiber. It has repeaters in different places in. As it goes up towards your brain, through your spinal cord, into your brain. And it's literally that wire got connected to some particular place in your brain, and that's how you know that it comes from that source and not another. Now, it'll be very interesting if we manage to make direct neural connections of things, that one will end up being able to remap all that stuff, and it'll be very bizarre. And I don't know how quickly one will learn. I have a feeling one will learn rather quickly. Oh, this is the sensation that comes with the thing is trying to tell one that the answer is seven or something. But I don't know quite what that will be like. And El Colombia is asking, do we need a lot more wires for vision compared to listening? Yeah. The optic nerve has about 10 million fibers. The auditory nerve has about 50,000 fibers. So there's a lot more data coming from vision is. What is that factor? It's a factor of 200 difference. So there's. And in a sense, that's reflected in computers when we say, well, 44 kilohertz is, you know, let's say 44. Let's say 100 kilobits per second. Let's say 100,000 bits per second is kind of what we need for a good audio. Good computer audio. But if we want good computer video, we'll need. Let's see, that will need to be at least 100 million bits per second or more. It's actually pretty much, now that I think about it, it's pretty much in exactly the ratio of the number of nerve fibers that we have going from our ears to our brains and our eyes to our brains. So, yes, that's how it works. It's very direct. It's very, in a sense, very straightforward. Okay, I probably. Okay, William is asking, I can't resist this question. What are the implications regarding the ability for the brain to acquire another sense, unlike the five senses we already have? You know, different organisms have different funky senses. You know, bats use echolocation. They make ultrasonic pulses where they. Which bounce off things. And that's how they know when those pulses come back. That's how they know how far away things are. And I don't know, what is it? The star mole, I think, is famous for having. I don't think it has eyes, but I think it has electrical sensors that tell it sort of where things are. And there are all kinds of different sensors that different organisms have. I think snakes have heat sensors. I think. Well, pigeons, for example, probably have magnetic field sensors in their brains. Maybe we have magnetic field sensors in our brains. We don't know for sure. But there are different types of sensing devices that exist in different organisms. And the question is, you know, different organisms take their sensory data more or less seriously. Like dogs, for example, have huge emphasis on smell as a sensory. As a sense, whereas we don't. And I think they smell as big. You know, vision is not so big for dogs. Vision is big for cats. You know, these kinds of things. And you can kind of see that reflected in the number of nerve fibers that come from those different sensory organs into brains. So the question is, if we were to have. If we were to introduce completely new sensors. Well, in effect, we have that all the time. Because any kind of measuring device is kind of giving us. Maybe our readout is because we can see it, you know, the meter in our eyes. Or maybe we have something that's, you know, clicking in our ear or something like this. You know, that's the way that we have extended our sensory domain is with all sorts of measuring devices. Now, when one can do that not through the medium of some existing sort of sense like vision or sound, but one can do it directly into the brain, what's that going to feel like? My guess is. Well, I don't know. My guess is not going to feel like much. I don't think it's going to be that different. I think people. It's like people who practice various kinds of things, like various kinds of meditation and things like that, think that they can start sort of sensing things with their brains that most people don't sense, like blood pressure, things like that, which most people are just not. The only way you can know your blood pressure is by measuring it with a blood pressure cuff. But, you know, there are people who, who say, well, I can learn to get that. I think actually sometimes even without something as fancy as that, you can learn some kinds of things because you can say, I'm going to have the sensor that shows me something about. I think, gosh, I think maybe even with blood pressure you can do this. But if you, if you have something, although it's hard to measure that continuously, so maybe, maybe other things. But, you know, you can say, I'm trying to learn how to sense this aspect of myself. Well, if I can get this display that shows me this aspect of myself, the, you know, can you, can you kind of learn what, can you correlate what you can see on the display with how you feel? And then, you know that that's a. That's a way in which you can kind of. You can learn by correlating what you see on the display with how it feels in your brain. And you can end up sensing some other thing like, like, I don't know, sensing glucose level in your blood is one thing that, you know, potentially you can learn. And potentially, now that there are continuous glucose monitors, you could potentially learn. Oh, you know, this feels like, you know, high glucose level or whatever. You know, I just ate something. I can feel the glucose peak from eating that potato or whatever else it is, and, you know, which people usually cannot. But something you can potentially learn if you have that feedback from an external source. All right, we should probably wrap up here for today, but a very genomic day and biological day today, and thank you for asking different kinds of questions, and I look forward to doing this again another day. Thanks very much and bye for now. Sam.