The Evolution of Sex

This has long been an interest of mine. It dates back to the old talk.origins days, prompted by a Creationist taunt with familiar tone – “I’d like to see someone explain the evolution of sex …” (with the implicit “hurr, hurr”). I articulated some thoughts, then was rounded on by the ‘mainstream’ community. I got a flavour of the world through Creationist eyes – an equally familiar tone: some very sharply expressed contempt and an invitation to f*** off back to high school and learn meiosis.

But, the more I thought about it, the more convinced I became that the ‘twofold cost’ picture traditionally presented, one that demanded an offsetting benefit of similar weight, was simply incorrect, however lofty the figures proposing it. The problem as expressed for a dioecious population (one with separate males and females: the ‘twofold cost of males’) or from the perspective of a locus on a genome (the ‘twofold cost of meiosis’) leads to the same result: an apparent halving of reproductive output for a female, or of survival odds for an allele. Yet since dioecy is quite rare, and ‘selfish alleles’ only exist when recombinant sex does, these cannot be taken as relevant to the origin of sex, and not all that relevant beyond it. The sensible perspective, it seems to me, is that of the haploid genome. From such a perspective, we are binary organisms – temporary unions of haploid genomes. It cannot be inherently more costly to unite then separate than simply to do nothing. The view that sex (as syngamy: gamete fusion) should have existed for only a moment before being erased by permanent diploidy seems wrong. Meiosis is the brief return of the native organism – the haploid. Permanent diploidy is cancer: a trap.

It’s more complicated than that …

I have written the above essay summarising my views. It is a bit of a dry and technical read – few pictures, even fewer jokes! This venue is not particularly conducive to a sensible discussion of the subject. It may be of interest to no-one: too amateurish for the professionals, too technical for the interested layman. Almost none of it is really original thought, being more a synthesis than anything groundbreaking. Beyond the choice of perspective, there are only a couple of ideas I have not seen elsewhere (I might remain coy about which those are!).

I’d prefer comments to be held off until the paper has been read (a big ask; it’s 50 pages!) but I hope it doesn’t descend into yet another discussion about semantics, metaphysics, the evidence for common descent, or any other of the conversations we have all had on a few dozen other threads.

My own time defending this is limited – I’m off to the Pyrenees for 5 weeks on June 24th. But, if anyone is interested …

237 thoughts on “The Evolution of Sex

  1. One thought that strikes me. If Maynard Smith is right and multicellular dieocious species suffer a real twofold cost due to production of males, the composite or cryptic twofold benefit typically conjured up to counter it must exist in isogamous populations too. They are supercharged, twofold-plus, all that surplus benefit kicking asexual offshoots with one hand tied behind its back. Just wait till dioecy comes along … this is not what seems to happen.

  2. John Harshman,

    And the reason sex isn’t abandoned more often is that its benefits generally outweigh its costs. That asexuality has benefits is shown by the frequent appearance of asexual species, and that it has costs is shown by the frequent extinction of most of them.

    Well yes, all naively true. My critique is aimed at the reasonableness or otherwise of what we should consider a true cost, in the opposition of the two sides in this one-way switch of genetic system.

  3. Allan Miller: My critique is aimed at the reasonableness or otherwise of what we should consider a true cost, in the opposition of the two sides in this one-way switch of genetic system.

    ?

  4. John Harshman,

    ?

    I think the twofold cost(s), the cost of recombination and the cost of segregation are illusory, errors arising from taking the diploid stance.

    It really might help if you followed the progressive argument I present in the essay. I realise you may have seen nothing in my words that would inspire you to do so.

  5. But as to the somewhat salad-y sentence, the costs are always presented relative to a competing asexual. The traffic is always one way – sexual diploid species beget asexual diploid species, never the other way round. I think the whole enterprise of trying to treat this as if it were an adaptive switch – as if it were two alleles at the same locus competing in one population – is flawed.

  6. Allan Miller:
    But as to the somewhat salad-y sentence, the costs are always presented relative to a competing asexual. The traffic is always one way – sexual diploid species beget asexual diploid species, never the other way round. I think the whole enterprise of trying to treat this as if it were an adaptive switch – as if it were two alleles at the same locus competing in one population – is flawed.

    Is there anyone treating it as an adaptive switch? Well, it certainly is within species. A great many species are asexual most of the time and only resort to sex in rare situations. How do you deal with those? It seems that in fact there’s a continuum.

    And I wouldn’t be absolutely sure that obligate asexuals never give rise to sexual species until I had surveyed all of life.

  7. John Harshman,

    Is there anyone treating it as an adaptive switch?

    Implicitly, yes. If one talks of the cost of a peacock’s tail, one is looking at the competing alleles increasing and decreasing its size, all within the population. Talk of ‘the cost of sex’ is very much of the same character. Not everyone is as blind to the much more complex and subtle dynamics of this boundary as I imply, but many seem to be. I think framing it in terms of cost-benefit analysis would be better dispensed with altogether, though no-one is going to listen to me … how do you evaluate the benefit of ‘being varied’? Or ‘being better equipped for a sexual ecology’? Or ‘allele optimisation at subgenome level’?

    A great many species are asexual most of the time and only resort to sex in rare situations. How do you deal with those?

    Those are still sexual species. There is no major conceptual distinction between a species such as ourselves which performs a series of mitoses in the one body, and aphids that extend those mitoses through a series of bodies. The haploids are still released, and therefore these are sexual species, because that is how genes flow. The distinction – hard to spot in practice, easy to see in principle – is between species that release haploids occasionally – genes can flow – and those that never will. The latter have speciated, but one might only realise with hindsight.

    And I wouldn’t be absolutely sure that obligate asexuals never give rise to sexual species until I had surveyed all of life.

    I think there are strong conceptual difficulties with this, though I would have to change my tune if there were a real counterexample. All covered in the essay (54 pages, y’know? I try to be thorough!).

  8. Allan Miller,

    Implicitly, yes.

    Scratch that, explicitly yes. Maynard Smith, Evolutionary Genetics page something-or-other:

    Consider an organism such as a herring, with equal numbers of males and females, and no parental care. In females, a gene A suppresses meiosis, and causes the development of diploid eggs that develop without fertilization into females genetically identical to the parent. Figure 12.5 shows that, when rare, such a gene would double in frequency in each generation. This result has been expressed by saying that there is a ‘twofold cost of sex’, arising from the needless production of males. It is clearer, however, to take a ‘gene’s eye view’: a gene A that suppresses meiosis is certain to be transmitted to all the eggs produced by a female, whereas a gene a that permits meiosis is transmitted to only half.

    Futuyma takes a similar line. Maynard Smith was no fool, and understood levels of selection, so I’m surprised to see him go for that approach.

  9. Mung,

    Sadly, sex is a topic of little interest to me.

    Hey, if I’d known that I wouldn’t have bothered!

  10. Just to labour my ‘anti-cost’ stance:

    Imagine a haploid population consisting of single-chromosome organisms fully capable of mitosis – haploid asexuals. Now, among them a mutation causes bearers to initiate occasional acts of syngamy. This is shortly followed by reduction back to haploidy.

    When paired, one would not say that the transient diploid they form suffers ‘genetic dilution’ – a cost of meiosis from halving. Genetically, the diploid’s just a pair of haploids, nothing more. They just go their separate ways. There are neither males nor anisogamy, so we can forget those too. There’s no recombination, so the only thing left is a notional ‘cost of segregation’ – breaking heterozygotes beneficial to the diploid state. But since those heterozygotes are formed by syngamy, reduction is hardly a net cost of the cycle.

    There’s no nontrivial ‘cost of sex’ at all at this point.

    Now imagine adding a single round of diploid mitosis before the reduction step. The diploid has acquired a little more permanence. Are there any new costs of sex? Not obviously, no. The two haploid genomes have gained a copy each, just as if they had each done a haploid mitosis, then out they come.

    So now we add a second diploid mitosis. Then a third, then a fourth … Does this change things? Not against the ancestral mitosing haploids it doesn’t.

    But of course given diploid mitosis, we have taken a step that may lead to diploid asexuality. So we suppose one lineage does this. Has sex suddenly become costly, due to this new mode? Not obviously. The diploid phase of the sexual population is in competition with the permanent diploids, which might be ‘saving’ reduction costs and segregational effects. But it’s hard to see where there is any inherent cost to remaining cyclically haploid against this competitor. This permanent diploid line would not even exist but for sex, so such lineages certainly cannot root the ‘mystery of sex’. They cannot get their genes into the sexual population, so the ‘gene’s eye view’ is an inappropriate stance also.

    We can take a similar approach with anisogamy. If the twofold cost does not appear with a tiny increment of asymmetry in the cytoplasm afforded to gametes in a bimodal population (dioecy), it’s hard to see where it appears with each successive increment. One can go all the way from a tiny difference to extreme egg/sperm type asymmetry; at no point does ‘the cost’ suddenly become twofold, but nor does it approach that value in proportional increments.

    It only appears, paradoxically, when you start with a dioecious population.

  11. Hi Allan,

    Thanks for your thoughtful manuscript; It has prompted me to come out of lurking here. Your “haploid perspective” is interesting and I agree it has merit in understanding the origin of sex, but I don’t understand why you take issue with the cost of males in dioecious species.

    Reproductive fitness in a sexual population is determined by the ecologically dominant phase. In your imaginary ancestral population there is no real difference between haploids and diploids, so your thought experiment of viewing diploids as temporary fusions of haploid cells works (though cells engaging in syngamy are wasting precious time not dividing). But the spread of genetic variants in organisms with a dominant diploid phase is largely dependent on their effect on the diploid phenotype, as all resources are acquired in that phase. Males do not invest resources in production of offspring, so they are wasteful. The millions of gametes they (that is, we) produce may be fantastic from the haploid perspective, but is not going to result in any long-term evolutionary benefits.
    I note that you use ecology as an argument in your manuscript against models that treat asexuality as an “adaptive switch” in genetic models:

    An ecological competition takes place, which should not be portrayed as if it were a genetic one..

    But since the sexual and asexual species are likely to be occupying the same niche, the outcome will be the same: The frequency of “asexual” mutations is expected to rise, as the proportion of asexual individuals goes up at the expense of sexual individuals. The fact that there is no common gene pool is irrelevant, as the asexuals will displace the sexual population altogether.

  12. BTW: All that talk of bdelloid rotifers reminded me of something: I had a colleague studying this very topic, who maintained that it has never been conclusively demonstrated that asexual lineages have higher extinction rates. The main argument for this is the “twiggyness”of phylogenetic trees, where asexual lineages are located at the tips. This is thought to demonstrate that these are young lineages. However, she showed that other factors, such as low transition rates to stable asexuality, or cryptic ‘speciation’ in asexuals, can generate such patterns as well. Published here

  13. Welcome to TSZ, biocv. Your first comments were held in moderation automatically. Any other comments you decide to make will appear immediately.

  14. biocv,

    Hi biocv Corneel, thanks for your feedback.

    I don’t understand why you take issue with the cost of males in dioecious species. […] Males do not invest resources in production of offspring, so they are wasteful.

    Speak for yourself! 😉

    I have significant difficulty in persuading biologists to drop an organism-based stance and view the transaction ‘from the inside’. Partly it’s my expositional skills, partly it’s their natural prejudice!

    I think my fundamental answer is in my preceding post. If one envisages an incremental progression from cyclic symmetry to cyclic asymmetry, either temporal (more diploid mitoses) or lateral (anisogamy), it is not clear at which point any cost becomes ‘twofold’, or even grows with the increments. And yet, if you start right at the end … there it is! That is rather bizarre, and calls into question how real that cost is for any entity that can do anything about it.

    In a dioecious sexual population, my haploid genomes spend equal average time in both genders. Genes that favour one state are selected in it, while silent in the other. Other than that, ‘from the inside’ the genetic transaction is indistinguishable from that in those early syngamising haploids I go on about. Competition is almost entirely realised in the huge and persistent diploid phase, but still haploids go in, haploids come out. There is no central genetic cost of dioecy that can be cashed by mutation to asexuality. If anisogamy is unstable, the expected future is isogamy, not asexuality.

    The millions of gametes they (that is, we) produce may be fantastic from the haploid perspective, but is not going to result in any long-term evolutionary benefits.

    That’s true of anything. However hard they try, on a replacement basis survival is cut back to 2 haploids or 1 diploid per diploid. But try they do, arms-race-style.

    But since the sexual and asexual species are likely to be occupying the same niche, the outcome will be the same:

    It’s important how you model that. Simplistically, you do the equivalent of taking a mixture of sexuals and asexuals and just giving them a good old shake. The next one to die or give birth is drawn at random because … well, because that’s what you do in a panmictic sexual population. But only one of these subpopulations has any reason to disperse from, or be drawn to, other members. Sex stirs the sexual population. In the ‘all-in-a-bag’ model you just stir the lot. I liken this to deliberately metastasing cancer and then deciding living individuals are a mystery.

    Of course ecology stirs too – resources attract individuals. But, the sexual is already there. It has variation. It can deal better with the wider ecology, much of it sexual by the time multicellular dioecy comes on the scene. It does not suffer from gene-conversion-induced homozygosity, imprisoned transposons, Muller’s Ratchet, hitchhiking effects. It is tuned for its role, which includes dioecy. Male gametes are a vector of dispersal, and males equally valid carriers and protectors of genes. And, the sexual population has faster anagenesis, and cladogenesis. If there is divergence, the asexual cannot beat them both with its clonal genome.

    I don’t feel the cost of males is really a cost of sex, but (if anything) a cost of dioecy. But I’d go further; it’s not a cost of anything. What is implicit in the argument is that sex suddenly had to start trying harder when dioecy came along. It could get along without a twofold benefit till then.

    I think the basic problem has been framed wrongly. ‘Here’s a (two-)generational 2fold cost, now come up with a benefit in the same terms”. But the situation is dynamic and complex, and not reducible to the same terms. Sexuals and their asexual offshoots are not parachuted in afresh to a virgin world. The dynamic situation preceding the arrival of any clone, and subsequent to it, informs its fate. You don’t give the one a simple selection coefficient half that of the other and then sit back and watch the carnage. The more simplistic the world one builds, the less effective sex is. But it didn’t start there.

  15. Allan Miller: Speak for yourself!

    Hey, I have two boys so I know about costs.

    Allan Miller: If one envisages an incremental progression from cyclic symmetry to cyclic asymmetry, […] it is not clear at which point any cost becomes ‘twofold’, or even grows with the increments.

    The fact that anisogamy and dioecy can evolve within a sexual system does not preclude that they entail some major deleterious effects at the population level. The costs of sexual conflict can be substantial, for example, but are only apparant at the population level, not the gene level.

    Allan Miller: And, the sexual population has faster anagenesis, and cladogenesis. If there is divergence, the asexual cannot beat them both with its clonal genome.

    Apart from the fact that species concepts get a little fuzzy when sex is abolished; What is preventing an asexual population from diverging? There is no longer any continuous mixing of genomes that is counteracting the diversifying selection pressure. It is sex that is an impediment to speciation so I don’t see why sexuals would have faster cladogenesis.

    Allan Miller: Sexuals and their asexual offshoots are not parachuted in afresh to a virgin world.

    Agreed. It is very hard to establish asexuality in a sexual world, and most species will be locked into the sexual reproductive mode, whether it is beneficial or not.

  16. Corneel,

    The fact that anisogamy and dioecy can evolve within a sexual system does not preclude that they entail some major deleterious effects at the population level. The costs of sexual conflict can be substantial, for example, but are only apparant at the population level, not the gene level.

    Yep, that’s true, but I do think it’s important not to just lump asexuals and sexuals into the one ‘population’. ‘The population’ in sexual models is a set of individuals through which genes can flow. Treatments of asexuality just chuck them in the same bin.

    It ‘costs’ male peacocks to have an elaborate tail, and so it must pay its way. What this means is that elaborate-tail alleles diffuse through this population, somewhat independently of all other loci. The losses of such alleles due to cost are balanced by their gains due to benefits. Were it not for the benefits, the alleles would diffuse out of the population, or antagonistic ones diffuse in. This is OK, you’re comparing apples with apples.

    There is no such situation with sex/asex. There aren’t two simple alleles, ‘Sex-‘ and “Sex+”, diffusing in the one population. There is, in the one case, an indivisible lump, with a composite selection coefficient because its loci are never exposed, and in the other, mosaic genomes each of whose loci is already variant and which can respond more effectively to selection, and process novel variants without interference. We can’t find any given genome in the sexual population that is necessarily fitter than the clonal genome, but we can be certain that the sexual population will have, somewhere, ‘the better allele’ for every single one of the comparable loci in the asexual – even though, at some loci, both will have the best available.

    I really think it important to resist the temptation to think of them as roughly the same, simply because they are constitutionally indistinguishable. The profound differences resulting from their genetic systems, despite their apparent identity as individuals, make their competition very unequal, which ‘twofold cost therefore twofold benefit’ completely obscures.

    Allan Miller: And, the sexual population has faster anagenesis, and cladogenesis. If there is divergence, the asexual cannot beat them both with its clonal genome.

    Corneel: Apart from the fact that species concepts get a little fuzzy when sex is abolished; What is preventing an asexual population from diverging?

    Well, their anagenesis is likely to be slower because of Hill-Robertson interference, gene conversion, dominance effects. But even if they diverged at the same rate as the sexuals, they wouldn’t be going in the same direction, because they are genetically uncoupled. My point is that we are invited to suppose that the asexual, merely by producing two-for-one, is overwhelmingly expected to supplant the sexual from which it arose (though only if dioecious). In a genetically static competition, that may happen. But if that sexual is undergoing anagenesis, or more clearly cladogenesis, then in order to extinguish the sexual ancestor’s modern descendants, it has to get rid of both. But it can’t be fully competitive with both of them at the same time. And somehow, it has to become coextensive with the parental range(s), despite not having any interest in other members of ‘the population(s)’.

    Allan Miller: Sexuals and their asexual offshoots are not parachuted in afresh to a virgin world.

    Corneel: Agreed. It is very hard to establish asexuality in a sexual world, and most species will be locked into the sexual reproductive mode, whether it is beneficial or not.

    Well … if it’s hard to establish asexuality, then I’m not sure by what measure sex is ‘not beneficial’ …

    I just find it odd to say that sex itself – rather than a particular variant of it – is costly for a population. If an asexual mutant has not arisen, being sexual is hardly an issue; it’s just doing what it’s always done: diploidy then haploidy. We kind of assume the relevant mutation is available ad lib, which is not a particularly justifiable stance either. Small populations are most at risk, but least likely to experience the mutation. Meantime, if the sexual system has become particularly elaborate – and I would cite dioecy as one such situation – it is not a generally simple matter to mutate to asexuality [eta – unless you’re a plant!].

    My objection is to this generalised idea that there is a universally applicable ‘twofold cost’ to sex, irrespective of all other considerations, and that this is equatable to a ‘twofold cost of meiosis’. And that, because of these erroneous equations, the entirety of sex is a mystery. As a consequence of the dynamics of temporary haploid pairings, I don’t find it at all mysterious, and don’t feel there is a need for any low-level ‘twofold benefit’, for which I fear we will be waiting forever. I think we already know as much as we are going to about the dynamics of these systems. It’s just a question of putting it together in the right order!

  17. OK, looks like interest might have fizzled out. I guess this idea will die with me then!

    I remain puzzled.

    No-one can actually say why genome dilution should matter to haploid genomes. The diploid, the level assumed to suffer this ‘cost of meiosis’, is not even a thing, evolutionarily speaking. It has no genes of its own; it has two haploid genomes.

    No-one can actually point to the bit in the transition from isogamy to dioecious anisogamy where sex becomes costly – at all, let alone twofold. It is reckoned costly by reference to the competition with assumed secondary asexuals, derived from the sexual line. But, if two-gender systems suffer a true population cost, why is asexuality the expectation, rather than a central tendency towards isogamy, or hermaphrodity? You still get sex, just not dioecy.

    In fact, anisogamy, despite apparent asymmetry, is genetically symmetrical. Gradual asymmetries in the ‘orbits’ of the two gamete types are complementary. The system doesn’t blow up just because it is possible for a large-gamete-producer to mutate to asexuality. It is a stable symmetry, resistant to both invasion and competition.

    I do think this approach largely eliminates the ‘mystery of sex’ as naively understood. One can write the twofold cost argument on a postage stamp. The counterargument took me about 10 years of thinking, research and 50 pages of fairly technical prose to get down. My difficulty has always been in persuading people to make the effort to follow it. It’s related to gene-centrism; Dawkins had similar difficulties getting people to see the merits of that viewpoint. I am, of course, no Dawkins.

  18. Allan Miller: It ‘costs’ male peacocks to have an elaborate tail, and so it must pay its way. What this means is that elaborate-tail alleles diffuse through this population, somewhat independently of all other loci.

    Ai, that is a poor example. The evolution of sexual ornaments depends critically on its gametic phase disequilibrium with genes affecting female preference. But I see what you mean.

    Allan Miller: There aren’t two simple alleles, ‘Sex-‘ and “Sex+”, diffusing in the one population.

    I think that is somewhat similar to the study of ecological speciation, where we are also looking at the tension between the cohesive effect of sexual reproduction and the need to specialise into specific ecological niches. Sometimes the benefits of certain genetic changes pay off an a higher biological level (ecology), or a different time scale (adaptive ability). Big deal.So why resist taking the reproductive system into account in genetic models, when it can clearly be such an important factor?

    Allan Miller: But if that sexual is undergoing anagenesis, or more clearly cladogenesis, then in order to extinguish the sexual ancestor’s modern descendants, it has to get rid of both. But it can’t be fully competitive with both of them at the same time.

    Let’s turn the tables here. “That sexual” is not just not mixing with the sexual population, it is not mixing among itself either. Every single asexual individual is starting its own separate lineage. How is the sexual population going to eradicate all those asexuals, when constant influx of malaptive alleles is constantly preventing it from adapting to any single specific environment? Sex can really be a nuissance!

    Allan Miller: My objection is to this generalised idea that there is a universally applicable ‘twofold cost’ to sex, irrespective of all other considerations, and that this is equatable to a ‘twofold cost of meiosis’. And that, because of these erroneous equations, the entirety of sex is a mystery.

    I’d agree with that. But even in your hypothetical isogamic garden of Eden, sex is a messy complicated exercise. Why waste time on syngamy when you should be multiplying as fast as you can? Sex is definitely in need of some explanation, is how I see it.

  19. Allan Miller: OK, looks like interest might have fizzled out.

    Heh, crossposting. Yes, the topic seems to garner somewhat less interest than the usual fare here. Even though the evolution of sex is a much harder nut to crack than that of “molecular machines”.

  20. Spoke too soon!

    Corneel,

    Big deal.So why resist taking the reproductive system into account in genetic models, when it can clearly be such an important factor?

    Sure, it can be an important factor. Indeed, that is kind of my point. The reproductive system is ignored in the models I’m criticising, where everything is just ‘a population’ – confusing ecological and genetic populations, as I attempt to illustrate in my manuscript. Despite the common belief that ‘producing twice as many grandchildren’ is an adaptation, it is better to say that the genetic system has impacts upon adaptation, which differ significantly. In particular, the environment into which the asexual is dropped always includes a variant sexual ancestor, and a wider sexual clade.

    There is clearly an ecological dynamic between the asexual and its ancestral population. But I am strongly resistant to the idea that this should lead us inevitably to the conclusion that these competitions MUST favour the asexual. That is the implication – wherever dioecy arises, it should have been killed off by asexual ‘females’, barring some strange magic not required at the isogamous stage.

    Allan Miller: But if that sexual is undergoing anagenesis, or more clearly cladogenesis, then in order to extinguish the sexual ancestor’s modern descendants, it has to get rid of both. But it can’t be fully competitive with both of them at the same time.

    Corneel: Let’s turn the tables here. “That [a]sexual” is not just not mixing with the sexual population, it is not mixing among itself either. Every single asexual individual is starting its own separate lineage. How is the sexual population going to eradicate all those asexuals, when constant influx of malaptive alleles is constantly preventing it from adapting to any single specific environment?

    It is not particularly necessary for the sexual population to eliminate the asexuals. It is only necessary, on ‘mystery of sex’ thinking, that the sexual population is not, in the general case, eliminated. We have this ingrained implicit belief that the tree of life should not contain any dioecious sexual populations but for a cryptic, unspecified advantage to sex. But if the populations are diverging, their contests are less and less relevant to their futures. It’s not sexuals that get rid of asexuals over evolutionary time, but the inherent difficulties of perennial diploidy. Competition doesn’t help, but more significantly the asexual lineages suffer genetic degradation due to gene conversion, inter alia. It is not necessary for the sexual cousin population to eliminate them by direct ecological competition to arrive at the pattern we see. They are battling everything in the wider ecology, from the back foot.

    I’d agree with that. But even in your hypothetical isogamic garden of Eden, sex is a messy complicated exercise. Why waste time on syngamy when you should be multiplying as fast as you can?

    I think you are exporting prokaryote expectations to the eukaryotic world. Eukaryotes discovered substantial leisure. They take much longer to replicate their genomes, and have longer interphases. In my ‘isogamete Eden’, it only requires that some quite small benefit (certainly not twofold) accrue from the diploid phase, and the act is paid for.

    Sex is definitely in need of some explanation, is how I see it.

    I agree, but the explanation is rather diffuse. There is a complex dynamic relating to the ecological situation at the moment a particular change arrives, and the genetic futures of the two competing populations one imagines conducting a contest at each stage. But there is certainly nothing new required, I think, just a reappraisal of assumptions.

  21. Allan Miller: That is the implication – wherever dioecy arises, it should have been killed off by asexual ‘females’, barring some strange magic not required at the isogamous stage.

    True, it is not the dioecy per se but the asymmetry in parental investment that makes the cost of males. In a system where only one sex nurtures zygotes, it is beneficial to switch to a system that only requires that sex. As a demonstration that males and females do not invest equally, consider this: In hermaphrodites, there exist some insane sexual conflicts where each partner tries to assume the male role, and force the other to take the female role. did you know about flatworm penis fencing? All of this evolves after sexual reproduction has already been established of course.

    Allan Miller: Competition doesn’t help, but more significantly the asexual lineages suffer genetic degradation due to gene conversion, inter alia.

    As remarked before, it has never been established that secondary asexual lineages are evolutionary short-lived. They may be performing as well as their sexual progenitors.

    Allan Miller: Eukaryotes discovered substantial leisure.

    And where did that come from? Are we talking about bottlenecks and genetic drift here? Does sex need an adaptive explanation at all then?

  22. Nitpick. page 49:

    …model flatworm C. Elegans…

    Caenorhabditis elegans is not a flatworm, but a nematode.

  23. Corneel,

    True, it is not the dioecy per se but the asymmetry in parental investment that makes the cost of males.

    Aha! we’ve pursued the twofold cost from isogamy (where it is not manifest, even as a ‘cost of dilution’), through anisogamy (where is not manifest in hermaphrodites) and have finally cornered it in dioecy with significant asymmetry of investment! That’s a very small, albeit self-absorbed, part of the Eukaryote Tree.

    Of course, that’s pretty much where Maynard Smith et al seemed to start, but in trying to get there from ‘the other end’, from an isogamous start, it’s hard to reconcile the two because as I say, there is no increment of anisogamy or of investment that clearly leads to an increase in cost provided that the asymmetries between the dimorphic gametes or individuals balance out over the generations. Which 50% residence time in each gender gives each gene. The response to any such cost would be expected to be to reduce a damaging asymmetry, not to favour the asexual switch. The asexual switch is only favoured if the asexual can thereby replace the sexual, which is by no means a given.

    Start from the asymmetric situation, and there’s a twofold cost. Start from isogamy, hard to see where it cuts in. That’s fishy.

    In a system where only one sex nurtures zygotes, it is beneficial to switch to a system that only requires that sex.

    So the question remains: beneficial for what? Genes aren’t bothered. ‘Lazy’ genes gain their benefit in the male, ‘parasites’ on those genes that provide care in the female. There are all sorts of tensions between the genes, as we see in conflicts over the foetus etc. But, both sets of genes reside in both sexes. Any given gene simply has its net selection coefficient split into two components.

    But what I still don’t get (to put my Columbo mac back on) is why we expect asexuality to emerge from such ‘conflict’ situations, rather than a balanced set of genes some with antagonistic effects, or the return towards isogamy from any tentative step towards asymmetry.

    As remarked before, it has never been established that secondary asexual lineages are evolutionary short-lived. They may be performing as well as their sexual progenitors.

    I think the ease with which meiosis genes can be detected in them speaks of evolutionary youth. If they were long term asexual, it is unlikely that these would be as readily detectable as they are.

    Allan Miller: Eukaryotes discovered substantial leisure.

    Corneel: And where did that come from? Are we talking about bottlenecks and genetic drift here?

    No. Eukaryotes are about 50,000 times larger than prokaryotes. They can eat, rather than having to diffuse materials. They don’t seem in quite the mad rush to replicate that prokaryotes are – in the “S*R” product of the fitness equation, they have shifted the balance towards S, and I think that’s what initially favoured diploidy too.

    Does sex need an adaptive explanation at all then?

    I guess it depends on the way adaptation is framed. Within a sexual population, locus-level alleles increase or decrease, and the population adapts. Within the asexual, entire genomes increase or decrease likewise. But transitions of the boundary itself are tricky to frame in the same way. Adaptation implies some kind of two-way shifting equilibrium. The boundary is (I think) 1-way, sexual to asexual. Sex isn’t an adaptation for a sexual organism. It’s doing it because its ancestors did, and it’s stable – it’s still doing it because it is resistant to invasion, from without or within.

    We certainly need an adaptive explanation at the origin. We need an explanation for syngamy, which is presumably given by an advantage, which I only hint at, of diploid union. We need an explanation for reduction, which I suggest might be provided by the mechanistic difficulty of keeping them glued together. Much of the rest, I think, is provided by the central robustness of the sexual cycle, and its many consequences. It is impossible for a gene to subvert from within, because as soon as it flips to asex, it pops out of existence as a recombining genetic locus, creating a competitive clone. Fancifully, that’s the one lever a selfish gene can’t pull, the one marked ‘self-destruct’. It can only be a selfish gene in a recombining system.

  24. Corneel:
    Nitpick. page 49:

    …model flatworm C. Elegans…

    Caenorhabditis elegansis not a flatworm, but a nematode.

    Well, same thing. Just like a whale is really just a cow, but we started calling them something different once they began swimming; and breathing out of the back of their neck.

  25. Allan Miller: That’s a very small, albeit self-absorbed, part of the Eukaryote Tree.

    Haha, fair enough.

    Allan Miller: But what I still don’t get (to put my Columbo mac back on) is why we expect asexuality to emerge from such ‘conflict’ situations

    Because, once a species has become dioecious, developmental constraints prevent a return to isogamy, but asexual reproduction remains possible in some lineages.

    Allan Miller: I think the ease with which meiosis genes can be detected in them speaks of evolutionary youth.

    I have never heard of that. Which study is that? Anyway, depending on your defintion of “evolutionary young / old”, there also are some examples of ancient asexual lineages, in addition to darwinulid ostracods and the already mentioned bdelloid rotifers.

    Allan Miller: They don’t seem in quite the mad rush to replicate that prokaryotes are

    It’s been a while since I have used life history theory, but I believe age at first reproduction is still a major fitness determinant, also in eukaryotes.

  26. phoodoo: Well, same thing. Just like a whale is really just a cow

    I believe the differences between flatworms and roundworms are even a bit larger than that.

  27. Corneel,

    Allan Miller: But what I still don’t get (to put my Columbo mac back on) is why we expect asexuality to emerge from such ‘conflict’ situations

    Corneel: Because, once a species has become dioecious, developmental constraints prevent a return to isogamy, but asexual reproduction remains possible in some lineages.

    Hmmm … I’m not really sure why developmental constraints prevent that return, but I can’t invoke similar to prevent a mutation to asexuality! What I would expect is that an allele that causes an increase in asymmetry that is detrimental would not become fixed in the population, a necessary precondition for an asexual to come along and reap the reward of a competition that avoids that cost. And, the challenges of the asexual applicable to the isogamous situation remain to be overcome. This would lead one to think successful replacement quite a rare event – not necessarily a nonexistent one.

    Allan Miller: I think the ease with which meiosis genes can be detected in them speaks of evolutionary youth.

    Corneel: I have never heard of that. Which study is that?

    Speijera, D., Lukešb, J., & Eliáš, M. (2015). Sex is a ubiquitous, ancient, and inherent attribute of eukaryotic life. PNAS, 112(29), 8827-8834.

    It’s the same study I use to support my contention that sex is foundational to the modern eukaryotic clade, including its asexuals. John Logsdon’s lab has also done work, for example the distribution of Spo11 that initiates DSBs in meiosis. Either their asexual status is actually wrong, and they are doing it, or they are young. (Or possibly there is exaptation, but it’s quite a universal pattern).

    Anyway, depending on your defintion of “evolutionary young / old”, there also are some examples of ancient asexual lineages, in addition to darwinulid ostracods and the already mentioned bdelloid rotifers.

    For sure. What we are looking for is an explanation for the pattern. It’s not necessarily all-sexual, any more than it’s justifiably all-asexual, as naive 2fold cost supposition would have it. No particular exception destroys the entire line of reasoning, nor can we say that its persistence is definitively due to 2fold cost. The tree is, in echo of Douglas Adams, ‘mostly sexual’. Although less so in unicellular forms, which lack ‘developmental’ constraints so probably find the switch mechanistically easier.

    Allan Miller: They don’t seem in quite the mad rush to replicate that prokaryotes are

    Corneel: It’s been a while since I have used life history theory, but I believe age at first reproduction is still a major fitness determinant, also in eukaryotes.

    Really depends what it’s exchanged for. If delay causes an increase in offspring survival, for instance. Or, as in the proposal on diploidy, a period that enhances the overall fitness of the haploid partners. Some of us stick around for years; it’s not all about fecundity.

  28. Allan, sorry to have been absent from this discussion. It’s been a busy time, and as the Evolution2017 meetings are coming up, it will continue to be busy for more than the next week.

    I have printed out your manuscript and will look at it soon.

    I do agree that isogamous species have no cost of “sex” (of outcrossing). The interesting issue would be, how to make a combined model of anisogamy and outcrossing. In a monoecious species, have three genotypes: one reproduces asexually, one is isogamous and outcrossing, and one is somewhat anisogamous and outcrossing. Could the latter spread? It would seem to depend on what the viability of various combinations of gametes whose masses were x, 1-x,, and 1/2.

    Apologies if this has already been covered here.

  29. Allan Miller: What I would expect is that an allele that causes an increase in asymmetry that is detrimental would not become fixed in the population

    Ah, but the increase in asymmetry is most likely not driven by detrimental alleles. I assume that the increase in asymmetry is driven by the increase in size of the female gamete, which would increase the survival of the zygote. The detriment is caused by the males’ inability to match this provisioning, as it would impair the motility of the male gametes. It is only the exiting of the sexual mode of reproduction which would relieve this problem.

    Allan Miller: Really depends what it’s exchanged for.

    Now I am confused. I thought you were saying that it is not inherently more costly to fuse and separate then to do nothing. Yet now you seem to be claiming that there needs to be a trade off to compensate for the delay in reproduction. And make no mistake; delaying reproduction is a costly thing to do, and a genotype that indulges in this will be displaced.

    The paper looks interesting. Thanks for the link.

  30. Joe Felsenstein,

    Hi Joe, thanks in advance – as you know, I’m off on an extended vacation soon, so no rush!

    My treatment is actually very short on models – almost entirely textual.

    The kind of isolated and static model you suggest would suffer from the generalised problem I try and highlight though – I argue that the more complex and real you make the situation, the more asexuality is penalised relative to its sexual relatives. A ‘big picture’ model would be computationally heavy, but necessary to give this 3-way competition realistic weights. It’s as much the dynamic of all the rest of the genome, and the interaction with the population(s) and ecology, that is under scrutiny during the progress of a secondarily asexual clone. But, I realise it’s necessary to make the problem tractable.

    I also wonder whether the normal path might be actually anisogamy in hermaphrodites first, which then gives rise to dioecy, a 2-step process?

  31. Corneel,

    Ah, but the increase in asymmetry is most likely not driven by detrimental alleles. I assume that the increase in asymmetry is driven by the increase in size of the female gamete, which would increase the survival of the zygote. The detriment is caused by the males’ inability to match this provisioning, as it would impair the motility of the male gametes. It is only the exiting of the sexual mode of reproduction which would relieve this problem.

    No, I disagree. If genes, when in males, cause sperm/pollen with poor dispersal, they’re detrimental. They will be outcompeted by genes that don’t.

    Another aspect to this is the role of males and their gametes in population stirring, and dispersal. In their pursuit of females, males range far and wide, and/or their numerous light/motile gametes do. This stirring underlies the panmixis of models. Now of course the asexual female genes don’t need to find anyone. But their spread around the population they are supposed to supplant is inhibited.

    Allan Miller: Really depends what it’s exchanged for.

    Corneel: Now I am confused. I thought you were saying that it is not inherently more costly to fuse and separate then to do nothing. Yet now you seem to be claiming that there needs to be a trade off to compensate for the delay in reproduction.

    Well, that was by comparison with a haploid that really was ‘doing nothing’, not one that was replicating! It was just to illustrate the difficulty of justifying the diploid stance for viewing the cycle: a hypothetical neutral case.

    There certainly needs to be an advantage to syngamy, otherwise it wouldn’t happen. But if we have a population of mitotically competent haploids not currently replicating, they are presumably feeding or trying not to be eaten – equally significant activities. While small, they are at risk of predation, and are limited in the size of what they can feed on. Syngamy provides instant increase in size to the benefit of both genomes. The alleles in the two also undergo complementation – heterosis. Any or all of these may be sufficient to increase the incidence of syngamy in the proto-population, even if there is some lengthening of the time between replication events for bearers of the capacity. The benefit need only be enough to offset that, in the first instance.

    And make no mistake; delaying reproduction is a costly thing to do, and a genotype that indulges in this will be displaced.

    Without offset, yes. But evidently, given that eukaryotes do delay cellular reproduction, some of them to extremes, there are many counterbalancing factors.

  32. *rank amateur butts in*
    I think I get the idea of “isogamy” as a precursor of differentiated sex,where the gametes cannot be described as male or female, but what is “syngamy”?

  33. Allan Miller: Another aspect to this is the role of males and their gametes in population stirring, and dispersal. In their pursuit of females, males range far and wide, and/or their numerous light/motile gametes do. This stirring underlies the panmixis of models. Now of course the asexual female genes don’t need to find anyone. But their spread around the population they are supposed to supplant is inhibited.

    That assumes two things that are certainly false in many species: 1) that males are the primary dispersers rather than females and 2) that tendency to dispersal is something both innate and sex-dependent. Now in fact in most passerines, females tend to move farther from their hatching sites than males. And a lot of dispersal is triggered by effects of crowding in many species.

  34. Joe Felsenstein,
    OK, (and thanks also to John Harshman)
    *emboldened to ask further daft questions*
    So the fusion event results in a diploid zygote? So sex has to be diploid? Diploidy results from sex or can haploids indulge too?

  35. Alan Fox,

    So the fusion event results in a diploid zygote? So sex has to be diploid? Diploidy results from sex or can haploids indulge too?

    Sex (according to me) is the cycle – alternation of haploidy and diploidy. Syngamy/fusion of two haploids makes the diploid, reduction/meiosis releases them.

  36. John Harshman,

    That assumes two things that are certainly false in many species: 1) that males are the primary dispersers rather than females

    I am trying to reinstate the male from its lowly perch. Of course it’s not just males that disperse and mix given sex, it’s mating in general. But this tendency to view males as a useless contrivance gives them insufficient credit. Whether they attract females or seek them, or just squirt their gametes out hopefully, they are involved in a ‘stirring’ that is not present in asexuals, besides being perfectly respectable containers for genes.

    and 2) that tendency to dispersal is something both innate and sex-dependent.

    Due to the need to find mates, some component of it must be in obligate species. I’m not saying that’s the only cause of dispersal.

  37. Allan Miller,

    You should be packing sun cream and bear repellent!

    Art Hunt convinced me that plants get overlooked when evolution gets discussed. Sex as a desperate measure when things get tough. And even then self-fertilisation as a last resort.

  38. Alan Fox,

    You should be packing sun cream and bear repellent!

    🙂

    Art Hunt convinced me that plants get overlooked when evolution gets discussed.

    That’s true. I try and keep them in mind. Fungi too; they are multicellular haploids that become diploid only briefly. And as for protists … everything that’s not a prokaryote is a protist, to a reasonable approximation.

    Sex as a desperate measure when things get tough. And even then self-fertilisation as a last resort.

    Diplocentricity. The idea it’s all ‘for’ the diploid organism. That’s what I’m arguing against!

  39. Allan Miller: I am trying to reinstate the male from its lowly perch. Of course it’s not just males that disperse and mix given sex, it’s mating in general. But this tendency to view males as a useless contrivance gives them insufficient credit. Whether they attract females or seek them, or just squirt their gametes out hopefully, they are involved in a ‘stirring’ that is not present in asexuals, besides being perfectly respectable containers for genes.

    Who considers males to be useless contrivances? What tendency? There’s no need to rescue anyone here, but I merely point out that your most recent attempt at rescue doesn’t make a case for males. None of this helps advance any point about either the origin or the maintenance of sex.

    Due to the need to find mates, some component of it [genetic impetus toward dispersal] must be in obligate species. I’m not saying that’s the only cause of dispersal.

    It isn’t true either. True, some dispersal is necessary in an obligate outcrosser, if only the dispersal of sperm from one individual to an immediately adjacent one. But that isn’t what you were talking about: it was panmixis. Now in fact no species approaches panmixis very closely, but you are implying that long-distance male dispersal is necessary to produce some approximation of it. Why approximate panmixis should be necessary, why it would have to be males doing the dispersing, and why dispersing needs any genetically programmed component are not at all clear. I’m trying to point out that perhaps none of that is true. Perhaps you wrote off the cuff and have lost focus a bit?

  40. phoodoo: Here we go again.

    Indeed. It was destined to happen! It is so ingrained that they don’t even know they are doing it.

    There must be an advantage to birds losing an ovary else it would never have happened! There must be an advantage to sightlessness else blindness would never happen!

  41. Allan Miller: No, I disagree. If genes, when in males, cause sperm/pollen with poor dispersal, they’re detrimental.

    I believe that anisogamy developed in multicellular species, as the division of labour enabled the provisioning of the female gamete, right? It follows that all alleles contributing to anisogamy were benificial, as they increased the survival of the zygote by stacking it with resources. The constraint on male gametes to remain motile is what prevents the re-establishment of isogamy, and allows secondary asexuals to reap a substantial advantage, once they arise.

  42. phoodoo: Here we go again.

    Are you saying it would happen without an advantage? Or are you saying the designer made it without it accomplishing some function that is advantageous to the organism? Are you saying syngamy is technically useless and only there because the designer sorta liked to make it?

  43. Mung: Indeed. It was destined to happen! It is so ingrained that they don’t even know they are doing it.

    There must be an advantage to birds losing an ovary else it would never have happened! There must be an advantage to sightlessness else blindness would never happen!

    Ironically, in the other thread where we are arguing about GAs and other forms of evolutionary computation, a paper has been cited that would help you make sense of why Allan said what he did. The general idea is that more complex traits usually only evolve if they (and depending on how complex they are, some of their intermediate stages) are adaptive.

    So, if we assume syngamy evolved, given that it is a rather complex trait, it and/or several of it’s intermediate stages must have been adaptive. And the challenge then is to identify in what way (under which circumstances) they could be adaptive.

Leave a Reply