Why aren't we all extinct?

About the strong harm of weakly harmful mutations and the future of mankind

Nadezhda Markina, PCR.news

To what extent does selection affect modern man, does selection compensate for the accumulation of unfavorable mutations? How to make the genome of the "ideal", more precisely, the most adapted fruit fly? Why is sexual reproduction necessary (and is it necessary at all)? Says evolutionary biologist Alexey Kondrashov.

Alexey Kondrashov is a professor at the University of Michigan in Ann Arbor, winner of the first megagrant competition, in 2010-2014 he led the laboratory of evolutionary genomics created by megagrant at the Faculty of Bioengineering and Bioinformatics of Moscow State University.

— Alexey, let's talk about the role of mildly harmful mutations and their accumulation in the body. The title of one of your articles contains the question: why haven't we all died out a hundred times already? Is it clear why?

— The fact is that the genetic burden, that is, a decrease in the average fitness of the population, is greater if the mutation is mildly harmful. Because if a mutation is very harmful, then it is crushed in the bud — negative selection never allows it to rise in frequency. And a mildly harmful mutation can reach a high frequency, a very mildly harmful one can even be fixed in the population. This was neatly shown by Kimura, Maruyama and Crowe in a famous 1963 article. But they were looking at one locus, and I thought that there are a lot of loci in the genome, and it turns out that the total load of these mutations can be prohibitive. I have, say, 100 million mutations in my genome, each of them reduces fitness by one hundred thousandth. Multiply 100 million by one hundred thousandth, we get 1000. So, I have to die 1000 times from the cumulative effect of these mutations, and why this does not happen, we do not know.

— What is the ratio of neutral and harmful mutations in humans?

— This is actually the same as the ratio of junk and working DNA. Because if a mutation gets into DNA that does nothing, it will be neutral, if it gets into DNA that does something, then most likely it will be harmful. There was a lot of talk on this topic, and the first work was done by my group in 2001. Then there was already a human genome, but there was no mouse genome yet. And we dug up 100 complete intergenic intervals that had already been sequenced in the mouse, compared them with human orthologous intergenic intervals, and looked at which fraction was aligned. Man and mouse are so far away that if there is no selection, then there will be no alignment. And so we estimated that 10% of all DNA is under the influence of negative selection. Which hinders evolution.

— Where did you work in America then?

— At the National Center for Biotechnological Information (NCBI). And a year after our work, the complete mouse genome was sequenced, and the main conclusion of the authors was that 5% of the genome is under the influence of negative selection, this conclusion was published on the front page of the Washington Post. They did exactly the same thing we did. But there was no link to us.

— You got 10%, and they got 5%.

— We had it right, but they didn't. Then our director of the institute, Colonel Lipman, a very colorful personality, such a "servant to the tsar, father to soldiers," or rather a servant to the president, father to scientists, burst into my office and began yelling that he would sue the authors of this article for not referring to us. In general, it is standard practice that the discoverer in modern genomic science is not the one who did, but the one who spent more money clarifying the third sign. And the current estimate is 8.4%, it appeared about five years ago. In the order of magnitude — 10% — there is no doubt. So the short answer is: 10% harmful, 90% neutral. Although nothing is completely neutral, but if one letter is replaced in the depth of some long intron, then the effect of this on my phenotype and my fitness will be negligible. And therefore, out of 100 mutations carried by a newborn, 90% do not affect anything, and 10% are harmful.

— And this 10% is an increase in our genetic load in one generation?

— It is not necessary to use this word here. The genetic burden is a relative decrease in fitness due to mutations, not the number of mutations. This is the Y-axis, not the X-axis. If each newborn carries 10 new harmful mutations, then in the absence of selection, the average number of harmful mutations per genotype increases by 10. If there is a selection, then less.

— Is negative selection able to compensate for the increase in the number of harmful mutations by 10 per generation?

— Either capable or not capable. If the population is in equilibrium, then it compensates. In accordance with the Haldane–Moeller principle, "one mutation is one genetic death" (genetic death is the inability of an individual to leave offspring. — PCR.NEWS). That is, the more mutations occur in one generation, the more harmful alleles must be removed to restore the average fitness of the population.

Suppose that in a population of flies there is some kind of distribution of individuals by the number of mutations, we assume that the average individual carries 1000 mutations. Let's construct a Poisson distribution. The standard deviation will be equal to the square root of this number — about 30. In each generation, due to the addition of the number of mutations, this distribution should shift to the right. We will remove from reproduction all those who have more than average mutations. Let's cut off the right half of this distribution. After that, the average value will shift to the left, by 80 percent of the standard deviation, that is, by 20. That is, by eliminating half of the individuals from reproduction, we can more than sufficiently compensate for the increment of mutations per generation.

Thus, selection in nature can compensate for the increase in the number of harmful mutations. And we know very poorly how selection works against harmful mutations in modern human populations.

The number of harmful mutations in the population is described by the Poisson distribution. In each generation, due to an increase in the number of mutations, the average value should shift to the right. But the elimination from reproduction of individuals with a number of mutations above the average reduces the average value.

"There are no labels on harmful mutations in the genome"

— How is it possible to study this case at all?

Well, you can use all sorts of clever methods to try to estimate the selection coefficients. My work with Lev Yampolsky and my son Fedya was on this topic 16 years ago, we roughly estimated the distribution of selection coefficients, considering possible amino acid substitutions. We built a histogram in the first approximation. And they showed that most of the selection coefficients live between 10-2 and 10-3. And it is not surprising, because if the coefficient is less than 10-4, then the selection is no longer effective, and if it is more than 10-2, then this is a very harmful allele. And then Shamil Syunyaev and his staff did it five or seven years ago on a lot of material, very carefully. And they confirmed that most of the amino acid substitutions in proteins lead to a drop in fitness in the range between 10-2 to 10-3.

— A mildly harmful mutation is if the replacement of a single amino acid slightly changes the conformation of a protein, but does not prevent it from functioning?

— Well, yes. Not necessarily the conformation, it changes something in the protein, quite a bit. Or it changes a lot, just protein is not important. We have a large margin of safety. Every healthy person carries, in my opinion, about 200 alleles with loss of gene function. And, in my opinion, ten of these alleles of loss of function are homozygous, that is, ten of my genes are completely broken, neither mom's nor dad's copy work. And nothing, I live. Naturally, if this gene is recessive lethal, then I will die. But a lot of genes are not recessive lethal, that is, you can break both copies and live. However, at the same time, the gene is still needed, because if the gene is not needed at all, if its breakdown does not reduce fitness at all, then this gene will die quite quickly at evolutionary times.

— Do modern molecular genetic methods help to understand this?

— What is the difference between the methods that exist now and the methods that were 30 years ago? The only difference is that now for $1,000 you can read the entire human genome. This is useful, of course. But there are no labels on harmful mutations in the genome. Let's say we take 100 people, sequence them and calculate a statistical norm in each place of the genome. Let's assume that the norm is something that has a high frequency in a given place. Rare alleles — they are most likely harmful. Harmful enough that selection does not allow them to reach high frequencies. Usually the frequency of the allele is either very large or very small. For two alleles to hang out with frequencies of about 50%, it is unlikely, this may be in a very small proportion of nucleotide sites. So, how many frequent variants in the human gene pool that are close to fixation, which are actually harmful? We have no idea. I believe there should be many millions of them, and I don't see any reason why there should be fewer. How will sequencing help here?

— Does the rate of accumulation of mildly harmful mutations depend on the size of the population?

— First of all, it depends on the mutation rate. Now it is believed that we know her well. Now only a lazy person does not measure the mutation rate by comparing mom, dad and offspring. And weakly harmful mutations depend on the size of the population, because a mutation is weakly harmful if selection is not effective against it. And selection is not effective if the selection coefficient is less than one divided by the effective population size. And what is very harmful in a large population will be weakly harmful in a small population. It will accumulate. In a small population, random drift, which knows nothing about fitness, chatters allele frequencies in all directions and interferes with selection.

But what is interesting is that the crown of creation, man, arose in a line in which the effective number was only 10 thousand. Monkeys have an effective population of 100 thousand, and when anthropoids appeared, it fell 10 times — simply because they are so big that there are not many of them in Africa. And this did not prevent one of the monkeys from turning into a human at all. And although Michael Lynch believes that selection is ineffective in small populations of large animals, I don't really believe it. The idea of a "drift barrier" seems doubtful to me, since the drift barrier did not prevent the evolutionary emergence of the human brain.

— Contrary to the layman's ideas, the mutation rate is no higher today than it was before, right?

— There are two points here. For humans, radiation has almost no effect on the number of mutations for the reason that it is very toxic. This drosophila fly can be given a monstrous dose, and it won't kill it, but it will cause a bunch of mutations. For mammals, radiation is so toxic that before you get a mutagenic dose, you will die. Therefore, radiation, even in the Chernobyl zone, as it turned out, causes very few mutations. Chemical mutagens are different, they can do a lot of things, but in modern life we do not consume them in such quantities. That is, such external mutagens do not significantly increase the mutation rate in the modern population. And then the only possibility is a change in demography. More definitely, there is a change in the age of men at which they have children, because mutations accumulate with the age of the father, it depends much less on the age of the mother. (The precursors of spermatozoa — spermatogonia — divide throughout life, and the greater the age of the father, the more mutations have arisen in them. At the same time, egg precursors, oocytes, are formed in embryonic development and then stored without passing through mitosis, so the age of the mother almost does not affect the number of de novo mutations, although it increases the likelihood of chromosomal abnormalities. — PCR.NEWS). More than 100 years ago, Weinberg noticed that children of elderly parents have more Mendelian pathologies. And he guessed that this was due to new mutations. Now it is clear that if the father is 60 years old, then his child carries four times more de novo mutations than if the father is 20 years old. Because in the first case, his sperm went through 600 divisions, and in the second case — only 150 divisions. Individually it works, but at the population level there is not such a big shift. So the mutation rate is now clearly not twice as high as it was in pithecanthropus, I can't vouch for one and a half.

How to turn off the selection of laboratory flies

— Tell us about your experimental work with fruit flies, in which you excluded selection in the population.

— It seems to me that 25 years ago we did it most elegantly, without using any tricks. We took 100 pairs of flies, put them in a test tube in pairs. From the offspring of each pair, only two were taken — a girl and a boy, all the descendants were mixed, again seated in pairs and also left two descendants from each pair. And so generation after generation. Selection is the differential reproduction of genotypes. And we forced all genotypes to have the same number of descendants, and thereby disabled selection. Then it turned out that fitness for one generation falls by 2-3%, and in a year when 30 generations have passed, fitness has more than halved. That's the whole science. At the same time, we spent the day and night in the laboratory, but it didn't require much intelligence.

— And how did you assess the decline in fitness?

— We forced them to compete with the third line in very difficult conditions. They took the original flies, took flies that had passed through many generations of such switched-off selection and forced both of them to compete with a certain test line in conditions of very high density, looked at who was better at fighting it. There was a difficulty there — to provide good control, for this we needed flies of the original population. We perverted ourselves in every possible way, including freezing embryos from the original population, then reviving them. We were advised by a specialist in cryobiology, Peter Steponkus. It is not difficult to freeze a fly, it is difficult to make it come to life later.

It was in the mid-90s — Svetlana Shabalina, Lev Yampolsky and me. The article was published in the Reports of the American Academy.

— So you modeled the accumulation of mildly harmful mutations in flies that reduced fitness?

- yes. We have modeled what happens in a population when selection is very much weakened — it falls apart very quickly. It is important here that the conditions of competition were tough, namely, in harsh conditions, mildly harmful mutations lead to a strong decrease in fitness, in mild conditions this does not happen.

— You are convinced that fitness is declining. And if this experiment went on and on, would they die?

— Probably, at some point they would simply stop reproducing. Harmful mutations cannot be accumulated indefinitely, and if pairs of flies do not produce offspring, there is nothing I can do about it. We recently did another experiment, repeating what Leonid Kaidanov did — we selected flies, on the contrary, for low fitness. Offspring from those parents who had low sexual activity were taken into reproduction, they mated slowly, produced a small number of offspring, and their development time was long. To get the next generation, the descendants of the worst flies were taken again. And for 15 generations, the population just died. We staged an "unnatural selection". This is Kaidanov's idea, but we implemented it in a freely crossing population, not in an inbred line. Of course, it would be interesting to sequence all this and understand what has accumulated there. But so far this work has not been completed. And it would also be very interesting to get the perfect genotype ... a person is impossible, but a fly is possible.

— What is the "ideal genotype"?

— Statistically ideal is a genotype in which there is a letter in each place, the most frequent in the population. A genotype free of all rare alleles. Since highly harmful alleles are always rare, it means that this will be a person free from highly harmful alleles. Such an average person. If it's uncomfortable to think about a person, let's think about a fly. It costs nothing to synthesize the genotype of this fly in a computer — take the genome of 100 flies, make alignment and consensus, and it will be the ideal fly you are looking for. But how to make this fly? So far, all the technology is not close to suitable for this. But it would be very interesting — such a fly, cleaned of all highly harmful alleles, will it be 5% more fun, stronger and more adaptable than an ordinary fly or twice?

This question was first asked by Herman Moeller in 1950. He, however, believed that every person is heterozygous for eight harmful alleles — the value is wrong by four orders of magnitude, the constant 8 came from somewhere else. And he believed that if there were no these eight harmful alleles in a person, then fitness would be 20% higher. In fact, we just don't know if he was right about that.

This is a purification from highly harmful alleles that are rare. And the second option is not a statistical norm, but a biological norm: let there be in each place of the genome what is best in this place. That is, if some mildly harmful mutation is recorded in the population, then we will also remove it. All people — or all flies, let's think about flies, not people — have glycine in this place, but in fact valine would be better. And here we have such an ideal fly, purified from both highly harmful and mildly harmful alleles, how much better will it be? We have no idea. I would give a lot to catch such a fly and study its properties.

— But one day it will be possible to do it?

— There are no laws of nature prohibiting this in principle. As experience shows, if laws do not prohibit something in principle, they do it sooner or later. But right now, all these Crispers aren't even close to that. So far, you can change one letter with Crisper and at the same time plant 20 errors. But a fly is still okay, it's more difficult with mammals. The fly has little garbage in the genome, all sorts of transposons, fly transposons are rare, they behave like highly harmful alleles. And in humans, half of his normal genome is dead transposons. To clean out all this filth from the genome — how much better will it get?

A very large proportion of large chromosomal rearrangements occur due to illegal recombination between repeats. For example, half of the cases of hemophilia A are recombination between two copies of a transposon that sit in the hemophilia locus close to each other, in the opposite orientation. And there is often something that gets stuck, recombines wrong, there is a restructuring in this place, and the gene simply breaks down. If at least one of these two transposons is removed from the human genome, the frequency of hemophilia A will decrease by half. There is no such time bomb in hemophilia B, all the mutations that lead to it are minor mistakes. And here half is such an illegal recombination. This is the cause of many hereditary diseases — "genomic diseases", or depome disorders, as James Lapsky called them. The term is unfortunate, all diseases are genomic, but it means that this is the result of macro-rearrangements in the genome. And most often it is illegal recombination due to close repetitions. These repetitions are usually not needed for nothing, it's garbage. Accordingly, if this garbage is thrown out of the genome, then there will simply be no such depome disorders.

— Do we live under selection today? Is selection completely absent in modern human populations?

— No, there is no selection. There is a concept of "opportunity for selection" — opportunity for selection. This is the variability in the number of offspring that we removed from flies when each pair brought exactly two offspring. In such a situation, there is no possibility for selection itself, because selection by definition is the differential contribution of different genotypes to reproduction. If each individual makes the same contribution to the next generation, there can be no selection. And if there is a spread, a dispersion of contributions of different individuals, then there is an opportunity for selection. This does not mean that there is selection itself, because it may turn out that this spread is purely random and has nothing to do with the genotype. So the first question is — is there an opportunity for selection in the modern human population? And it's very interesting here. What is the difference between the modern population and the normal one that has been throughout almost the entire history? The fact that thanks to medicine, poorly viable individuals survive and sometimes even reproduce.

The first is a decrease in child and adolescent mortality and the second is a drop in average fertility. Previously, 70% of children did not live to reproduce, but then each aunt brought 12 children, each uncle brought even more. The modern demographic transition, which began in Europe and is now taking place in equatorial Africa, includes two things: a decrease in child mortality and, with some delay, a decrease in fertility. Due to this delay, a catastrophic population growth has occurred, which is now ending, because the birth rate is falling all over the world. The decrease in mortality naturally leads to the fact that the opportunity for selection decreases. And a decrease in average fertility is not necessarily accompanied by a decrease in the spread in fertility. Here it is necessary to take not an absolute spread, but a relative one. Let's say earlier a woman gave birth to an average of 7-8-9 children, and now on average two children — 0-1-2-3. The absolute spread is now smaller, and the relative one is larger.

To say that selection does not work in modern human populations is incorrect. Selection for viability does not work because almost everyone survives. And there is quite a variation in fertility. Although the opportunity for selection is not the selection itself. I think that selection is greatly weakened due to cultural and social factors, not biological ones. Instead of going to a monastery, a barren aunt undergoes some kind of therapy, and produces someone. Accordingly, this weakens selection. Or some aunt produces a lot of children, not because she is very fit, but because her father told her so. In general, it is unknown. But these are all philistine considerations. There is such a quantitative concept — a selection differential. This is the difference between the average value of the trait after selection and the average value of the trait before selection. So, what is the breeding differential in terms of the number of mildly harmful mutations in the modern population, we have no idea.

There is also such a hidden factor as prenatal mortality, and it is 70%, because only 30% of conception leads to birth, in ideal conditions. It was shown more than 30 years ago, and everyone was very surprised then. We took well-off young Americans who want to have a baby and conducted a pregnancy test every three days. And it turned out that a huge number of pregnancies occur and immediately disappear in the first two weeks, when no one knows about them. Another thing is whether it is selection or something else - it is unclear.

When spontaneous abortion occurs, it is often a chromosomal anomaly. I don't understand why at all. Is it so difficult to make a normal meiosis? What nonsense! Some danio fish — 99% of its fertilized eggs will develop into fry. Although these eggs are in a Petri dish and no one protects them. And here is a human embryo — and the probability is only 30% that you will not die. There are all sorts of interesting evolutionary considerations on this topic, but no one knows how correct they are. The dry residue is that the possibility of selection before birth is very strong, because the mortality rate is very high. No one knows how much this mortality is associated with harmful mutations. Apparently connected.

— There is a point of view that in modern humanity there is a selection against the genes of intelligence. People who are genetically disposed to receive higher education, on average, have fewer children than those who do not have these genes. Is this a selection?

— It was shown quite recently. Of course, this is selection. Any connection between your genotype and contribution to the gene pool of the next generation is, by definition, selection. There are such data for the USA, where a correlation between fertility and education was shown. And in Iceland, a correlation was shown between alleles that contribute to education and fertility. The correlation is negative: the more alleles you have that contribute to education, the fewer children you have. That is, in fact, the level of education itself was thrown out of this connection and replaced by genes. But it's very interesting, yes. It is all the more interesting that when there was no education, conditional pithecanthropus had alleles that allow their distant descendants, that is, us, to receive higher education, although the pithecanthropus themselves did not receive any education. And now, when we get it, these alleles turn out to be harmful — they reduce fitness.

— If we estimate at what speed, with weakened selection, there will be a decrease in fitness, can we say how much humanity will still be enough?

— I don't think we are radically different from a fly here. So, if the selection is completely disabled, after 20-30 generations it will be very bad. Even earlier, probably — we still have a much bigger target for harmful mutations. But it is not a fact that selection is really completely disabled or even radically weakened. This is the most important question.

Why is sexual reproduction necessary?

— One of your articles concerns the evolutionary role of sexual reproduction. Is it probably related to curbing the accumulation of mildly harmful mutations?

— This is an amazing situation — no one knows anything about this topic. She was quite popular, but now everyone is somehow fed up, no one is engaged in the evolution of sexual reproduction. At the same time, the issue has not been resolved. I think I can show and explain why all the proposed mechanisms for the benefits of sexual reproduction do not work. Obviously, I'm wrong somewhere, because there is some mechanism, but I have no idea what it is.

Some role of sexual reproduction seems obvious — an increase in diversity, and the offspring do not show harmful mutations that parents have if they are recessive.

Hiding recessive harmful mutations in a heterozygote is not sexual reproduction, it is diploid. Diploidism and sexual reproduction are often confused. Sexual reproduction does not increase diversity by itself. It eliminates deviations from independence in the distribution of alleles of different loci. Let's think about haploids, such as some fungi or algae. When such a beast has sex, it forms a diploid with someone, and this diploid immediately goes into meiosis. All life takes place in the haploid phase, only the zygote is diploid.

— Do they benefit from recombination?

— They are obliged to benefit from recombination, because recombination and meiosis is a clear adaptation, it is such a complex and fragile thing that would not exist if it were not supported by selection. So, for general reasons, they certainly get the answer. But what this advantage is, we have no idea. What does sexual reproduction do? It randomizes the distribution of alleles at different loci, that is, it does this: what I will have at locus B (B or b) does not depend on what I have at locus A. Or mathematically , the frequency of the AB genotype will be equal to the product of the frequency of the A allele by the frequency of the B allele . And why is this randomization needed? Unclear. There is a huge science on this topic, but there is no answer there.

— Will he show up sooner or later?

— Probably, it will appear. There are not very many logical possibilities there, and they were all invented 30 years ago. But all these logical possibilities do not work, in the sense that they do not agree with the data. There is certainly a mistake somewhere in my reasoning, but I do not know where. Now everyone is arguing that these non—random associations, the destruction of which makes sexual reproduction, recombination profitable - that they are created due to genetic drift. In my opinion, four orders of magnitude are missing in this theory, this is nonsense of a gray mare and this cannot be. Or maybe I don't understand something, maybe the drift works there.

— It turns out that in evolutionary biology such fundamental issues as mildly harmful mutations and the role of sexual reproduction remain unresolved for so many years?

- yes.

— And yet modern molecular genetic methods somehow help to solve some of these issues at another level by looking into the genome?

— One obvious example is the measurement of the mutation rate. Because I still remember those times when it was discussed, the rate of mutations in humans per nucleotide is 10-9 or 10-8. And now everyone is measuring the third sign. That is, there has been a revolution in the measurement of the mutation process due to technology — simply because now you sequence the mother, the folder and the offspring, look at their genotypes and see all the mutations. This is an absolute benefit from modern methods. It is customary to scold all these GWAS, in fact they are also wonderful. When a huge number of individuals are taken and their phenotype is simply measured, which is relatively easy, and the genotype is sequenced. Something is visible there, yes. But there were no fundamental breakthroughs. It is clear that the model of mildly harmful mutations works — in the sense that all sorts of complex diseases are most likely associated with the cumulative effect of hundreds and thousands of alleles in the genotype. Each of them individually has a very small effect.

— Do they deal with theoretical problems of evolutionary biology in the world? As I understand it, there are few people doing this in our country, so your laboratory was unique.

— We were not engaged in theoretical problems, but in data analysis, one and a half people were engaged in theory. In Russia, modern evolutionary biology, except for graduates of my laboratory, in my opinion, simply no one is engaged. Egor Bazykin runs his laboratory at Skoltech. And there is no one else.

Now, in general, by the word "evolutionary theory" they mean not at all what they used to mean. Previously, evolutionary theory was, as they say, a direct task. There are some factors of evolution — mutations, selection, random drift, recombination — and there is a question: what will happen to the population if such and such factors act on it. This is a direct task. And now — on the contrary: we see what has become of the population, and we are trying to understand how it came to such a life, what factors affected it. This is the inverse problem. Now 95% of what is called evolutionary theory is the solution of the inverse problem. We see the genetic structures of the population and try to understand what kind of selection acted on it. They do this a lot, with varying success.

— If we turn to the evolution of the coronavirus, how do you think its future evolution is predictable? 

— Completely unpredictable. Who knows? In order to predict evolution, you need to know the adaptive landscape. It is necessary to know for all potential coronavirus genotypes what properties they will have. How do we know this? Can there be five more mutations that will lead to the fact that its mortality rate will increase from 1% to 25%? And even such that it will spread better? This is a nightmare scenario — an even more pathogenic option will arise and displace everyone. Delta is slightly more pathogenic, but the virus does not care about its pathogenicity, the virus is important how it spreads. Delta spreads easily. And won't the next variant, which will be much more pathogenic, spread? And who knows? I keep repeating that the virus does not have to become white and fluffy over time, there is no reason for this. So it may or may not be. (The interview took place even before omicron appeared. — PCR.NEWS.)

— The logic of such reasoning is that it is unprofitable for a virus to kill its host.

— This is nonsense. The virus kills its host by the time the host is no longer contagious. The virus is not a humanist or an anti-humanist, the virus is a dumb aggregate of molecules. The biology of the rabies virus is such that it must be lethal, otherwise the host will not start biting everyone and it will not be transmitted anywhere. And covid mainly spreads while a person is still at the very beginning of the infection, and what happens to him then, the virus does not care at all. So evolution can go either way. But where will it go? Will there be a higher virulence or a lower one? Don't know. This depends primarily on how the contagiousness and virulence of this pathogen are related. And until we know that, the greatest theorist here won't tell you anything.

"Being a 'Soviet scientist' is a below—average pleasure"

— What do you think about the future of evolutionary biology as a science?

— I don't know. It will somehow develop if everything is not covered with a copper basin. If everything goes as it goes, in 30-50 years it will be impossible to live humanly on Earth. A global environmental catastrophe is taking place. And you're asking me about evolutionary biology. The number one problem is climate change, the whole future of world civilization depends on it. So far, we are moving along a trajectory on which the Earth will soon be unsuitable for civilized life.

— Do you not believe in the intelligence of mankind?

— I don't know, maybe now humanity will come to its senses and stop deciding who owns Jerusalem, Crimea and other territories, and get down to business? But rather no than yes.

And yet, while the climate apocalypse has not yet come, tell me, how do you see the future of science in general? And in Russia, in particular?

If the current trend continues, it makes sense for novice scientists to leave here. That's all the foreseeable future of Russian science.

— Can science develop in an authoritarian or totalitarian state?

— Well, to some extent it can. I have been a "Soviet scientist" for more than ten years. Basically, science was meaningless, but there were exceptions of varying degrees of exclusivity. And if you look at the last ten years — something is happening here, something is fluttering.

— Some quite successful scientists believe that it is "difficult, but possible" to work in Russia.

— Competitive science is not being done in large numbers here. There was my laboratory in evolutionary biology, now there is Egor, and ten other people. Well, there are a thousand more such people in the world. I advise strong young people to leave if circumstances allow. Because being a "Soviet scientist" is a lower—than-average pleasure, and you will most likely do much less than under normal conditions. There will be no global successes on the path along which everything is rolling now. There will be some local ones — a genius will be born and prove some theorem. Hardly all of them will be taken out. But now Russia is two percent in terms of world science. And it will only get worse while what is happening is happening.

— Alexey, is your life in the village a kind of escape, including from big science?

— Rather, just the next stage of life. Grandpa is old... At 65, almost no one will do anything worthwhile in science anymore. I don't mean bossing, but real work. There are, of course, people who are engaged in science until they are very old or think they are. But it's better to leave on time. I was engaged in science at a very high level, and it is absolutely not available to me now. After all, no one plays chess at the highest level at the age of 65. No, I don't suffer from intellectual idleness — I attend Egor's seminar every week — once it was my seminar, and now Egor - and there is some benefit from me there.

— Here you are engaged in apple tree breeding in your village — but as a scientist, as an evolutionary biologist. Does it fascinate you?

— Actually, it's too late for me to do breeding, it's a long-term process. Just collecting interesting genotypes adapted to local conditions is already a great thing. And there was no horse lying around. The genetics of the apple tree is a very interesting thing, and it does not require a lot of intelligence, no offense to apple geneticists, be told. It's like if a person has been engaged in classical ballet all his life, and before retirement he goes into characteristic dances. It's the same thing. From classical ballet to characteristic dances around the apple tree. At least, it's nice.

Portal "Eternal youth" http://vechnayamolodost.ru