Why sex? Experiments on fruit flies suggest it evolved to resist infection

August 16, 2015 • 11:45 am

Of course there’s a proximate reason, at least in our species, to the question above: “Why sex?” That answer is fatuous but true: “Because it feels good.” Of course it feels good—because pleasurable orgasms and the desire to copulate are the evolutionary cues prompting us to leave our genes via mating.

But why mate with another individual in the first place? Why not, as do many species, simply reproduce asexually, so that if you’re a female you simply produce offspring from eggs that have not undergone the process of meiosis (i.e., reducing the genome by eliminating one of each pair of chromosomes, a genome restored to fullness when it united with a sperm)? Lots of species can do this, at least occasionally, often by producing unfertilized but viable eggs that have a full chromosome complement.

You can show that there is in fact a significant evolutionary loss caused by having sex, and by undergoing the characteristic processes of sex: recombination (chromosomes swapping bits with the other chromosome of a pair) and segregation (members of different chromosome pairs randomly assorting themselves into eggs or sperm). The loss is in fact two-fold (it’s called “the cost of sex” or “the cost of having males”), so that, all things equal, an individual that can reproduce asexually leaves twice as many copies of its genes as an individual that has sex. In other words, a gene for eliminating sex, and reproducing asexually, should sweep through populations. According to evolutionary theory, every species should reproduce asexually!

But that’s not the case. The vast majority of non-microbial species on Earth reproduce sexually. Given the cost of doing so, there must be some tremendous evolutionary advantage to having sex, one that is strong enough to outweigh the big twofold cost of having sex.

Sadly, evolutionists don’t know what that advantage is, so the evolutionary explanation for the ubiquity of sex is a mystery. It’s one of the great Black Boxes of my field.

One explanation, which has some support, is that having sex enables you to produce more genetically diverse offspring: diversity that is a result of recombination and segregation in the parental genomes. If you have more diverse offspring, perhaps some of them would have the right genetic combinations to withstand infections or other environmental challenges. But if all your offspring are just like you, as in most asexual reproduction, then there’s no chance for diversity. If you can’t resist an infection, neither can your kids. This could select for genes that produce sexual reproduction, allowing some offspring to survive.

This is called the “Red Queen hypothesis” for the evolution of sex, stemming from a scene in Through the Looking Glass in which Alice and the Red Queen are running, but always stay in the same place. The name is appropriate because, if the hypothesis is true, organisms are always running to stay ahead of parasitic infections; but no matter how diverse their offspring are, there will always be new parasites (or newly mutated parasites) that come along, and so the pressure for maintaining sexuality remains.

One can show that, under certain conditions, this increase in the diversity of offspring can select for sexual reproduction—that is, the advantage of having recombination and segregation, and in uniting your egg or sperm with a gamete from another individual, can outweigh the twofold cost of sex. The evidence for this, which is not extensive but still accumulating, is that there is an correlation among populations of some species between infection rate and sexuality. In snails in New Zealand, for example, populations that are infected with a worm that can sterilize them often reproduce sexually, while uninfected populations tend to be asexual. There’s other evidence as well, but it’s of this correlational type.

Now, however, a new paper in Science by Nadia Singh et al. (reference and free download below), suggests that one can see an advantage of one aspect of sex—recombination between members of chromosome pairs—as a response of single individuals within their lifetimes. Using fruit flies, they showed via a series of clever experiments that infected Drosophila produce a higher proportion of recombinant offspring than do uninfected individuals. This suggests not only that parasites might be a factor that selected for sexuality in their hosts, but that hosts can somehow “sense” that they’re infected, and produce more genetically diverse offspring. Having such an evolved “sensing” mechanism would of course be adaptive, allowing you to produce more diverse offspring—for if you’re infected, your offspring are likely to be as well.

Here are two individuals of Drosophila melanogaster, the study organism, having sex. The male, with the black abdomen, is to the right:


What Singh et al. used is an age-old technique in Drosophila: measuring crossing over (recombination) between members of a single chromosome pair (in their case, chromosome 2). This is in fact the way Alfred Sturtevant showed, back at the turn of the 20th century, that genes were lined up on chromosomes in a given order. He was a true genius, and here he is about the time he did that experiment (note the label, which I recall Sturtevant wrote himself):


In the case of Singh et al., the researchers measured crossing over between ebony and rough, two genes on the second chromosome that cross over about 20% of the time.  What they did is make females that had the mutant genes (affecting body color and eye texture respectively) on one chromosome, and the two “wild type” non-mutant genes on the other. Since ebony and rough are recessive forms of the genes, these doubly heterozygous females  have a normal appearance.  However, when you cross them to males homozygous for ebony and rough, you can measure the amount of recombination by the proportion of offspring showing only a single mutant trait (circled in the diagram below, which is from the paper). Normally we see about 20% of these offspring—the further genes are apart on the chromosome, the more crossing over between them will occur, and the more recombinants you see.

Screen Shot 2015-08-16 at 8.16.01 AM

It turns out that when the doubly heterozygous females (within dashed box to left above) are infected with the bacteria Serratia marcescens, they produce, when crossed to ebony rough males, significantly more recombinant offspring than do either noninfected or mock-infected females (i.e., those pricked with the injection needle, but with no bacteria injected).

Clearly, infection increased the degree of recombination—one aspect of sexual reproduction—although the p value, showing how likely these results could have been simply a random fluctuation and there was no real effect, was a bit higher than I like. It was 0.03, meaning that even if there was no effect of infection on recombination, the experimenters would observe an effect this large in one of 30 trials. The “cutoff level” is 0.05, so biologists consider p values below that to be “significant”. So these results were marginally significant, explaining why the paper was published in Science, but I’d like to see more experiments, particularly ones showing that the increase in recombination is genome-wide rather than affecting part of one chromosome.

It’s possible that, for reasons we don’t understand, infection simply causes breaks on chromosomes, or on the second chromosome in particular. One could test that as well by seeing if infection by other organisms, like tiny worms, would also have the same effect. Another experiment using injection of heat-killed bacteria showed no effect, so live bacteria are required to initiate the production of more-diverse progeny. More important, an experiment with wasps (see below) gave a similar result.

There are two other explanations  for the increase in the proportion of single-mutant offspring. One is mitotic recombination: that crossing over might occur not during sperm and egg formation (at the time when “sister chromosomes” pair up), but beforehand. Alternatively, there might be transmission distortion, so that there is no increase in the amount of crossing-over, but that the recombinant chromosomes somehow get preferentially sorted into the eggs, making it look as if there was an increase in crossing over. In all cases this could still represent an adaptive evolutionary response, for all three mechanisms yield a higher proportion of more diverse offspring. What we’re talking about here is simply the mechanism of getting more diverse offspring.

I won’t go into details, but you can test these alternative mechanisms by taking advantage of two features of Drosophila: males show no crossing over during sperm formation (we have no idea why this is!), and because all “normal” recombination occurs 4-5 days before fertilized eggs are laid. Ergo, if the appearance of recombinant offspring increases after this, it must be transmission distortion. It turns out that the increase in the number of recombinant offspring does indeed appear due to transmission distortion: recombinant chromosomes are preferentially put into the eggs. We have no idea how this is done.

Finally, the authors did one more experiment, injecting Drosophila larvae with parasitic wasps rather than bacteria (these wasps are a normal predator on fly larvae in nature). And, as with the bacteria, the female larvae that successfully fought off the wasp infections grew up (after pupating) to produce a higher proportion of recombinant offspring than control, uninfected larvae. Below is the figure showing the difference, which was even more significant (p = 0.0002) than with the bacteria. The proportion of recombinant progeny is on the left scale, and it’s lower in control than infected wasps:

Screen Shot 2015-08-16 at 9.11.48 AM
(From paper): Box plots illustrating the distribution of recombination fractions in D. melanogaster strain RAL73 in control and wasp-infected females. The median is marked with a black line; the first and third quartiles are rep- resented as lower and upper edges of the box, respectively. The whiskers extend to the most extreme data point no farther from the box than 1 times the interquartile range. Recombination fraction is shown estimated over the entire 12-day egg-laying period

The curious thing here is that these are larvae that are infected, and larvae haven’t yet developed the cells that produce eggs. Somehow the effect of being infected carries over from the larval to the adult stage.

This result, then, adds to the growing pile of data suggesting that, at least in some cases, the evolution of sex is connected with resistance to infectious agents—i.e., that sex is an adaptive response that wards off infections by producing so many diverse progeny that some of them will have combinations of genes that resist infection. Or, to say it yet another way, genes for sexual reproduction are advantageous because they happen to be present in those individuals that have other genes allowing them to better survive infections. Sex genes “hitchhike” on infection-resisting genes.

The paper of Singh et al., however, extends previous work because it shows that an adaptive response can evolve not just over generations, but can be a plastic one: that recombination can be increased during one’s lifetime when an infection is detected. (That response, of course, if it’s truly adaptive and not just an epiphenomenon, must also have evolved over generations.)

Is infection, then, a general selective force that produced sexuality in most species?

Who knows? Sexuality is nearly universal in multicellular organisms, but were all of them subject to strong selection by infectious agents that overcame the high cost of sexual reproduction? Of course once you develop the complicated apparatus of sexual reproduction, it’s hard to go back, so there’s a kind of inertia that can retain sexual reproduction even if it’s no longer advantageous. But, contra that, there are many cases of sexually-reproducing species having some asexual reproduction, so this reverse evolution can and does happen. Why don’t sexual organisms revert more often to asexual ones, given the advantage of asexuality? Is infection that pervasive, and such a strong selective force? Or are there other factors that select for sexual reproduction?

Here are some questions I have that would extend this paper, which—make no mistake about it—is very good.

  • Is infection-induced “recombination” (i.e., transmission distortion) in flies specific to the second chromosome? It would be nice to know whether recombination is increased throughout the genome, as one would expect were the Red Queen Hypothesis true. This would be an easy experiment to do.
  • Could you show the evolution of higher recombination in the laboratory? You could set up “population cages,” each containing a large number of genetically diverse flies, and then infect half the cages, leaving the other half as controls. After a year or so (about 25 generations), you might expect to see the evolution of higher recombination in the infected cages. It would be best to do sequential infection using several agents, as one would like to impose constant selection on the flies, and if you’re fully adapted to an infectious agent—so that it no longer harms you—then there’s no further selection for sexual reproduction.

One thing I should add at the end is that we’ve long known that artificial selection for increased recombination is highly successful in Drosophila. You can, for instance, take the kind of females shown above, and choose for further breeding the ones that produce a higher proportion of recombinant offspring. Over generations, this can lead to a pretty strong increase in recombination (you can also select for decreases). Further, you can also select for increased recombination between a specific pair of genes on a single chromosome, leaving the rest of the genome with no change. Clearly there are genes around with variants that can change the rate of recombination, which of course is a prerequisite if that aspect of sexuality evolved as an adaptation.

Alice and the Red Queen


Singh, N. D. et al.  2015. Fruit flies diversity their offspring in response to parasite infection. Science 349:747-750.

88 thoughts on “Why sex? Experiments on fruit flies suggest it evolved to resist infection

  1. Wasn’t it Alfred Sturtevant who as an undergraduate made the first genetic map (rather than Calvin Bridges)?
    Sturtevant, A. H. (1913). The linear arrangement of six sex-linked factors in Drosophila, as shown by their mode of association. Journal of Experimental Zoology, 14, 43–59.

      1. Sturtevant and Bridges are heroes of mine. Having spent a lot of time microdissecting polytene chromosomes, I have huge respect for Bridges’ efforts in mapping polytenes! While Painter recognised the utility of polytene chromosomes, it was Bridges that made the useful map.
        At the risk of nit-picking, ebony and rough are on chromosome 3!

  2. Very interesting paper AND blog post, Jerry. I remembered this paper from Levi Morran’s lab. @ Indiana University about somewhat similar experiments done in C. elegans. Could you comment briefly about how these two papers compare?


    Another thing I am curious about is whether recombination per se has been shown to increase fitness in individuals in a population. We know that inbreeding definitely leads to accumulation and expression of recessive traits, but has some work been done on the beneficial effects of recombination?

  3. Fascinating article and commentary! One of the most interesting posts ever. In a way it blurs the line between Lamarckian and genetic inheritance; acquired characteristics can permanently (not just temporarily as in methylation/epigenetics) affect the state of the lineage that descends from that individual.

    I wonder about something else, though. Is the cost of sex really 2x, even in hermaphrodites like snails? In a normal population of organisms, the vast majority of nucleotides are not polymorphic. So in fact after sex, each parent is still going to transmit some very large fraction of its genome to its offspring, even though some of those genes get there by a circuitous route (via the member of the opposite sex who inherited them from some common ancestor of both male and female, or via new mutations). So the cost of sex, for hermaphrodites at least, should be quite a bit less than 2x in most populations, unless I am missing something. In fact, the cost of sex is zero if all members of the population are genetically identical. Right?

    1. It depends on the hermaphrodite whether they experience a 2-fold cost of sex. The hermaphrodites that I know best (earthworms, some snails) must still mate with another individual and so these only pass on half of their genes. But I expect their are other hermaphrodites that can mate with themselves and pass on all of their genes. I cannot think of any example, however.

      1. But Mark, you’re missing my point. In most populations, most nucleotides aren’t polymorphic, so each individual is passing most of its genes into its offspring, because the contributions from the other sex mostly are identical-by-descent with the contributions from the first parent. There would be no cost at all if the population had no genetic diversity. The cost must be a decreasing function of genetic diversity, and should approach zero as diversity approaches zero. It should not be 2x.

        1. That’s logic which is analogous to group selection. The individual gene or allele can’t know whether the other parent will pass on an identical allele or a different one, so I don’t see how your argument works.

          1. It’s not a matter of “knowing”. Think about the limiting case of zero diversity. There is no cost to sex then. The cost is dependent on the diversity.

        2. @Lou
          Your problem seems to be related to an initial misconception of the cost of sex as the cost of meiosis or genome dilution by George Williams (1971. “Group Selection.” Aldine-Atherton; 1975. “Sex and Evolution.” Princeton Univ. Press).

          While it’s the most intuitive way to start thinking about the cost of sex as due to half the genes being “thrown away” in meiosis, this is irrelevant for the advantage of an asexual mutant that is also reproductively isolated from the sexual population it arose in. It’s advantage depends on it’s rate of increase in comparison and competition with the sexual individuals. It does not help an asexual mutant to transmit all its genes to all its offspring, if it does not also produce more offspring of equal fitness than the average sexual female.

          The standard all else equal assumption supposes that asexual mutants have as many and as fit offspring as sexual females, that the latter invest half their reproductive resources into sons and that males contribute nothing but genes to reproduction. That way you get the ominous twofold advantage of asexual mutants over sexual competitors.

          Treisman and Dawkins (1976. ‘The “cost of meiosis”: is there any?’ J. Theor. Biol. 63:479-484) were among the first to sort this out. John Maynard Smith suggested the alternative conception of the cost of sex as the cost of males in anisogamous species. That is, the cost does not lie in meiosis or genome dilution and is therefore also independent of the question whether most of them are polymorphic or not.

          P.S.: The cost of meiosis or genome dilution re-enters the picture, when you have asexual mutants that retain some way to re-introduce the asexuality alleles into the sexual parent population. For example, in hermaphrodites whose eggs develop into clonally identical offspring, but who retain their male functions and can thereby transmit the gene for asexual reproduction by mating with sexual individuals.

            1. Welcome,
              Stelzer has a paper in the current issue of PNAS (112: 8851) that would also be a good starting point and for following up references.

      2. “can mate with themselves and pass on all of their genes.”

        Republican candidates adamantly opposed to sex education.

    2. You seem to be talking about the entire genome. Fixed genes don’t enter into the equation, because they’re going to be there regardless. It’s the small percent of genes that vary which see their chances of entering the next generation halved due to sex.

      It’s rather related to the issue of relatedness. We’re 50% related to each of our full siblings. That doesn’t mean we share only 50% of our genes, though, because all humans shared more than 99% of our genes. The 50% figure is the percentage of variable genes that are shared. Or, put another way (perhaps a better way), there’s a 50% chance that any non-fixed gene you inherited is also present in a full sibling.

      In a hypothetical population where all genes are identical, there would be no cost to sex. There would also be no sex, because why involve someone else when you have all the information you need yourself? More importantly, no such population could ever exist. Even identical siblings will have some genetic differences due to mutation since their initially common embryo split.

      1. Yes, that is the point I am trying to make. The cost of sex is much lower than 2x, because most alleles are identical in both parents. The cost goes down to zero as genetic diversity decreases to zero.

      2. It’s the small percent of genes that vary which see their chances of entering the next generation halved due to sex.

        To be clear, it’s the gene’s chance of entering a particular offspring that’s halved by sex. A gene’s chance of entering the next generation can be pushed as close to unity as you like by having more offspring. With ten offspring, for instance, you’ve transmitted more than 99.9% of your (variable) genes.

          1. Though in some species, males are quite reduced in complexity, and in some they’re even consumed after consummation.

            1. I particularly enjoy contemplating the pleasant life of one of the deep sea fish. The male is just a minuscule little guppy that attaches to the much larger females body like a leach and thus is carried about serving as a sperm reservoir. Talk about your free rides! I think eventually he (the male fish so attached) is absorbed when the sperm is used up. It doesn’t get any better than that.

        1. Yes, but again, that chance is not halved by sex. Most of the time that gene will be in the offspring with or without sex, so for those genes sex has no cost.

          1. But by the same token, for those genes reproduction offers no benefit. They’re already fixed in the gene pool, so it doesn’t matter whether I pass them on or not. My selfish genes care about making it into the next generation only to the extent that they differ from everyone else’s genes. From a Darwinian perspective, the genes we all share are selectively neutral and don’t contribute to individual differences in fitness.

              1. (1000000*2)/1000000

                The fixed genes cancel out just like the million does. The remaining genes halve their chances of making it into the next generation by taking part in sex.

                No matter how complexly you try to analyze the machinations, that’s what it will reduce to.

  4. Is infection that pervasive, and such a strong selective force?

    I would think it is. It is certainly a strongly selective force. Think of how much metabolic energy is devoted to host defense.

    1. I’m inclined to view parasites and pathogens as just one facet of a variable selective landscape in which offspring may find themselves. Producing a clonal monoculture of offspring targets a specific narrow range of that landscape, while sexual recombination targets a broader area and therefore seems more likely to score an adaptive hit, even in the absence of infectious agents.

      1. I thought the same. So it would follow that those organisms that reproduce asexually are not subject to a lot of pathogens. There must be other forces beyond pathogens that would make asexual reproduction good enough. That kind of research would be interesting too.

      2. The Red Queen Hypothesis emphasizes that sex evolved to help us keep pace with microbial-type pathogens. Since these organisms evolve very fast due to having very short generation times, we sexuals must have a special means to ‘churn’ our genomes just to keep up with then. Mating ~ doubles the variation of our offspring, and before mating their is meiosis which includes crossing over and independent assortment of chromosomes. So the gametes that are made even before actual mating are all genetically different.

    2. How long do organisms without an immune system survive? Isn’t it generally weeks to months? That seems like a pretty pervasive threat.

  5. Maybe you should have slapped a “NSFW” warning on that pic of hawt Drosophila melanogaster sex. Thanks for the “male, with the black abdomen, is to the right” explanation, though. I was worried it was a strapped Drosophila chick pegging some poor Drosophila dude.

  6. You say that “The curious thing here is that these are larvae that are infected, and larvae haven’t yet developed the cells that produce eggs. ” But I am sure that they do have germ line cells since these first appear in embryos — they are then called ‘pole cells’. I do not know when Drosophila germ line cells enter meiosis in females, but in female mammals they begin very early in fetal development. It might be the case that female Drosophila begin meiosis as larvae.

  7. Very interesting post! I remember reading a bit about the Red Queen hypothesis in Dawkins’ earlier books, and Carl Zimmer’s Parasite Rex, and I think Matt Ridley talked about it in a book titled after the hypothesis.

  8. Reading on Sturtevant, a couple of things jumped out for me:

    “He loved solving all kinds of puzzles and saw genetics as a puzzle for him to decipher.”

    I can see the link between things like crossword and recombining DNA.

    “He had an impressive memory and composed and edited papers in his head before writing them down from memory.”

    Such remarkable talents are just amazing to contemplate – especially when I consider my own feebleness in that department. But, as they say, it’s all in your genes. You might say he had lucky genes. His grandfather and father were academics, and his son William became a renowned ethnologist.

    1. I hate how my brain can be slow to get things. I had a friend in school who was always about 3 seconds ahead of. Wake up lazy synapses!

  9. The titl”: “Why sex? Experiments on fruit flies suggest it evolved to resist infection” might suggest that sex originated to resist infection, while the experiment shows that infection resistance might be a factor in the maintenance of sex. Sex (meiosis and recombination) seems to be as old as the eukaryotic cell, and intimately linked with eukaryotes. That argues against infection as involved in the origin of meiosis.

    1. I wondered about that myself. Under what conditions did sex arise. What were the contemporary forces at work. It might be hard to say, but that would help to resolve the mystery.

    2. Sex (meiosis and recombination) seems to be as old as the eukaryotic cell, and intimately linked with eukaryotes. That argues against infection as involved in the origin of meiosis.

      Not at all. The selective impact of infection most decidedly precedes the origin of eukaryotes.

  10. The “Hot Dog” photo is Sturtevant in his World War I army uniform. He wrote the annotation.

    I remember seeing Sturtevant when he visited the Genetics Department at the University of Wisconsin for a time in 1965. At one point another visitor, Warren Ewens, was giving a seminar and making an argument that came to a different conclusion than one made by Sewall Wright. Wright, upset, got up and went to the front blackboard and wrote equations arguing against Ewens, and kept arguing, refusing to listen to Ewens’ replies. Sturtevant was seated next to Jim Crow. He leaned over and whispered into Jim’s ear “I wonder what it would be like if Wright argued with Wright?”

    The Crows also went with the Wrights and the Sturtevants to a park near Madison. Sewall Wright and Alfred Sturtevant, both in their 70s, decided to climb a rock outcrop. The Crows could only look on, horrified, as they struggled up it. Fortunately neither fell.

    (As for the sex-and-parasites theory, one argument against it is that sexual reproduction is so universal that it is hard to see that they effect is strong enough in all those species to counteract disadvantages of outcrossing.)

    1. I’m really sorry that I never got to meet Sturtevant, one of the early and true greats who lived into my era. By all accounts he was a really nice guy, and a genius.

      He had, of course, a famous falling-out with Dobzhansky, and nobody seems to know what it was all about (I have an early letter from Dobzhansky that refers to it).

      The real question that Sturt should have asked was this: “I wonder what it would be like if Wright argued with Wright? Who would be RIGHT?”

  11. I don’t get a chance to comment much these days Jerry, but I decided to take a moment to tell you how much I appreciate the science posts here at WEIT.

    Thanks for writing!

    1. But only if we are talking about already present selective pressures, natural selection is blind to the future.

      1. No, I think he has a point. Pressures are pressures. Parasitism is just one among many environmental forces. These may be already present, just as much as an infection. It need not be forward looking.

        1. But also, evolution is backward steering in the sense that progeny is infused with what worked primarily for the parent generation and to a weakening degree going back. There is a generational delay in the learning process of the genome, though I haven’t seen any model expressing it. Similarly fixation has a delay before it happens.

          These time constants should be accounted for when discussing selective pressures, at a guess. They should supply an ‘integration time’ under which they respond to forcing. I.e. pressures may need not be present each generation, but ‘dominantly’ (in some sense that need math to be defined) over several.

      2. There are microscopic organisms that do both – asexually reproduce in good times and sexually reproduce in bad times. It is IMO trivial to see how ‘already present selective pressures’ would evolve such a system: the bugs that – due to inherited tendencies – flip mechanisms at the optimal time produce more offspring than the bugs that don’t. Thus evolutionary pressures on current organisms create the appearance of forward thinking, though no actual forward thinking occurs.

  12. I am always mystified by why sexual reproduction seems so odd. Obviously I am missing something very fundamental but the following explanation seems so obvious that there must be a flaw in my thinking.

    Under asexual reproduction any chance improvement in a gene is likely to be repeated fairly faithfully down the generations. However, these improvements must happen by chance mutation and too many mutations cannot be allowed to occur at once for fear of a deadly variant appearing. By contrast, under sexual reproduction, each variation is determined not by chance mutation but by a choice between two alleles, each allele having been pre-tested for fitness for purpose within one or other parent and having survived down countless generations. In short, the more complex an individual the more traits need to be worked upon, all at once, to achieve an overall meaningful improvement within the shortest possible time span. By contrast, asexual reproduction is likely to succeed at just one thing, faithful reproduction and survival of the genes. Here, any chance improvement in any one of millions of traits is likely to have a tenuous expression in the likely survival outcome of the phenotype. One or two small improvements within many thousands of genes do not provide much of an edge in the game of life. This is perhaps why more asexual reproduction is seen in the simplest of plants and animals whilst sexual reproduction is the norm in the more complex life forms. In evolutionary history simple beings undergoing asexual reproduction came first but it was probably sex that spurred that essential boost to evolution that produced the vastly more complex lifeforms that we see today.

    I have a feeling that the theory of “The Selfish Gene” et al has been at the root of the difficulty for theoreticians to accept sex as an obvious choice for life forms to adopt. They have a problem with the gene’s “desire” to get into the next generation when set against the carving up suffered by the chromosome in the process. I have my own thoughts as to why this is not incompatible with “The Selfish Gene” theory but will hold back until, or if, I receive any explanation for the above mystery.

    1. I’ve always felt the same, though not able to express it as pithily and in as much detail. And I’ve felt the resistance to this line of thought had something to do with thinking that it involves selection at a higher level that just the gene. Which I gather is the meaning of your last sentence?

      1. Thank you Diane for your kind words and well done, you were right, I was actually referring to selection at a higher level, or something along those lines. The selection method I had in mind is so simple as to be almost invisible. Let’s just call it “Whatever Works”. This is the strategy whereby meiosis is passed down the generations and whatever different rules for the swapping of bits of the genome are followed, this would be the strategy that competes and the most successful one is that which is preserved. I have to admit that this is only a speculative explanation for “Whatever Works” and there are, no doubt, many others but my point is that there could, in some mysterious way, be a plan for sending genes into the next generation which acts at a higher level than that of the gene itself.

        I see in your reply to Scientifik (reply no. 17) that Jerry stated that the math just doesn’t work for arguments similar to ours. I have to admit that I am not sure what the math reveals but I have no doubt that the understanding of the mathematical argument would be somewhat above my pay grade!

    2. Admittedly sex recently has sailed up as probably at the root of (extant) eukaryotes, but it doesn’t make much difference in complexity. The average gene content of an eukaryote isn’t even an order of magnitude larger than the average prokaryote (~ 3-4 kgenes vs ~10-15 kgenes, IIRC.) Sure, the latter may have a lot more regulation mechanisms, but where they present at the root?

      [Since I mentioned Lane in another comment here, I may as well remind of his energy theory on eukaryotes. The main difference is not the nucleus but the mitochondria. The latter makes for simplified, effective, distributed energy plants that no prokaryote can mimic. The resulting boosted energy density, up to a factor 100 000 [!] enables that much more protein turnover, in turn enabling evolving larger genomes.]

      Conversely, sex seems to be concurrent with eukaryotes without being bound to complex multicellularity. That said, prokaryotes can’t survive completely asexually because of Muller’s ratchet, which you seem to describe. They too voraciously exchange pieces of DNA (and even RNA I think), and likely always has since the first cells could have been leaky and still work.

      I am only superficially familiar with the selfish gene model, but once I grokked it I think of it as I finally grokked evolution. Once seen, it can’t be unseen.

      Re the math, maybe Jerry is referring to the evolutionary loss he describes in the article, the ‘dilution’ of selfish genes. The problem then is to understand how the loss (which differs between explicit sex and gene transfer mechanisms of prokaryotes) is mitigated in sexual and incompletely “asexual” organisms.

  13. I have nothing useful to add. I just wanted to say this was fascinating, and to note that I do read the science stuff. Lots of us, I suspect, are like me and don’t feel able to comment on posts like this, but genuinely appreciate and respect the extra effort they take. Lack of comments is definitely not lack of interest!

    1. Glad you said that, Heather, because that was my reaction too! This is yet another subtle and insightful post that I will have to re-read before I can begin to understand it properly. Thank you so much, Jerry, for taking all this trouble to explain and summarise the issue and what the further implications are.

  14. “One explanation, which has some support, is that having sex enables you to produce more genetically diverse offspring: diversity that is a result of recombination and segregation in the parental genomes. If you have more diverse offspring, perhaps some of them would have the right genetic combinations to withstand infections or other environmental challenges. But if all your offspring are just like you, as in most asexual reproduction, then there’s no chance for diversity. If you can’t resist an infection, neither can your kids. This could select for genes that produce sexual reproduction, allowing some offspring to survive.”

    I have always considered it to be the best explanation. By reducing genetic diversity of the gene pool of a species, we decrease the ability of that species to adapt to changing environment. Another benefit is development of entirely new features.

    1. I once wrote essentially the reasoning in your last paragraph to Jerry, and he replied that the math just doesn’t work.

  15. I posted this to Twi**er, and got this reply from someone: “Looks like I have a new pick-up line.”

    I told him I didn’t think “It’ll stop your progeny getting infections,” went too well with candlelight and roses. 🙂

  16. There was a study a short while back on some closely related fishes (in Mexico, I believe), some of which had recently reverted to asexual reproduction. They checked the severity of parasitism over some time period (20 years?), and found that the sexually reproducing populations were much healthier.

    Points to the same conclusion from a different direction – sex probably evolved as a defense against parasitism, and remains for the same reason.

    I think it’s important to remember, too, that sex evolved in single-celled organisms. The functionality was there before multiple cells were, and the outwardly complicated machinery we see today to facilitate sharing genes developed over time from much simpler organisms. It was never the case that some large asexual multicellular organism started evolving sex to cope with parasitism.

  17. I thought Olivia Judson covered this in her NYT series of a few years ago fairly well, and while it didn’t answer the theoretic questions, she quoted a lot of relevant empirical data; cloners (above the microscopic level) get regularly wiped out. This includes large parthenogenetic species as well as the rare cases where males develop an adaptation that allows them to replace the mother’s genetic contribution entirely with their own (which happens…but not for long).

    So while the theory question isn’t answered, the empiricist in me says that whatever sex does, its pretty certain that it provides a net advantage over cloning. The “half my genes” problem is more than compensated by the “greater chance of survival” advantage.

    1. See my comment #4; I think the cost of sex is much much less than “half my genes”, since most of the genes the other parent contributes are also mine. We both got them from the same common ancestor, or from fortuitous mutations.

      1. Yes, Lou, but the argument that a gene that allows you to reproduce asexually will be present in far more copies the next generation than the alternative allele, regardless of relatedness or genic similarity, still works–I think!

        1. Thanks, yes, I think that’s true. For a given novel gene, the cost will still be 2x, and that’s probably what matters….But the cost for the whole genome will be much smaller than 2x.

          1. IIRC there are various fungi that actually flip between the two mechanisms; they clone when there’s lots of food/sun/room to grow but switch to sexual reproduction when environmental conditions are stressful. That would seem to me to again support the hypothesis that the genetic mixing provided by sex gives a survival advantage that more than compensates for its genetic cost…at least in conditions where the environment is not extremely stable.

            The existence of such organisms may also, however, support the notion that being able to switch is a big advantage, and the reason macroscopic life forms don’t do it is either because the multicellular eukaryotes that gave rise to animals got stuck on an evolutionary ‘low peak’ early on, or because keeping both mechanisms in place is hideously expensive.

            1. “…the notion that being able to switch is a big advantage, and the reason macroscopic life forms don’t do it…”

              Some insects do.

  18. Just like to add my thanks to those expressed above. I always seem to learn something interesting from your science posts.

  19. Great post, Dr. Coyne. My PhD advisor recently wrote a paper looking at this topic from a slightly different angle. He and his collaborators worked on several species in the genus Oenothera (Evening Primrose) that have evolved functional asexuality – the plants still reproduce sexually but not the chromosomes do not recombine or assort independently. Using comparative transcriptomics they found high rates of fixation of putatively deleterious mutations and a relaxation in the efficiency of selection in functionally asexual species compared to other closely related species that have normal recombination and assortment.

    If anybody is interested, here is the paper:


  20. Back in 1929 James Thurber and E.B. White wrote a book called “Is Sex Necessary?”
    Now we have the answer: “If you get infections, YES!”

  21. Fine post. At Noor’s course “Introd. …” the same/other points are given that can make sexual reproduction an advantageous option:
    Despite many costs of sex
    – Recombination can produce advantageous combinations of alleles (don’t have to wait until the right mutation(s) occur).
    – Recombination can accelerate adaptation.
    – Recombination allows the population to “unload” itself from bad mutations (stopping the “ratchet” effect).
    – Recombination may be particularly helpful in variable environments.
    And we can add now “resist infection” as another reason. The bad mutations and its accumulative effect (in asexual repro.) seems to me sufficient to justify sex repr. . Why is this not the case for the majority of living beings?

Leave a Reply