Principles of Evolution, Ecology and Behavior: Lecture 10 Transcript

Professor Stephen Stearns: Okay, today we're going to talk about evolutionary conflicts, and this is an area of evolutionary biology that contacts other disciplines, including the Humanities, in interesting ways. I was reading Nature this morning and there was a review of an off-Broadway production about the life of Robert Trivers, whose picture you're going to see here, and it was about Robert Trivers as a disturbed young genius at Harvard coming up with ideas of conflicts of interest and whatnot, and the actor on the stage was going on about how tormented the young Trivers was, and he was smoking pot and all of this stuff, and having his great ideas, and upsetting the faculty at Harvard, and finally giving up in disgust and going to Santa Cruz and meeting Huey Newton and the Black Panthers and all of that stuff. Okay? And there was a guy in the back of the theater who was chuckling and came up and congratulated the actor afterwards, and he said, "You got it just right." And it was Bob Trivers, who was watching. Bob's a professor at Harv--at Rutgers now.

So we're going to be talking about interesting stuff today. And I want you to be aware that the first part of it is well-founded, well-supported, experimental science, where the conclusions that I'll present to you are quite reliable. And at the end of it, I am going to go into some speculative stuff, where the conclusions aren't so reliable, but it's very interesting. And I want to say that up front so that--and I'll give you a signal when I transit from reliable stuff to speculative stuff, because I don't want you thinking that the speculative stuff is written in stone.

So let's begin by looking at these plant flowers. These are pPantago flowers. Plantago is something that you may very well have dug out of your lawn. It is a rosette plant. It is a common plant around the world, and it is gynodioecious, which means that it has two kinds of flowers. It's got flowers that have both male and female parts, and it's got flowers that just have the female parts and greatly reduced, or almost absent, male parts, male sterile parts. And evolutionary sex ratio theory tells us, in fact, that from the point of view of the nuclear genes, it's best to have a 50:50 sex ratio and to be investing 50% in male and female function; and we will come to that soon.

However, in this organism evidently, in some of them, this 50:50 sex ratio has been subverted, and all they do is reproduce as females. Now it turns out that the genes that control this morphological switch are in the cell organelles; they're not in the nuclear genome. The genes that are sitting in the organelles in the cell, be they mitochondria or chloroplasts, can only get into the next generation through female function, through the eggs. They are not transmitted through pollen. It is in their interests to take the organism that they're sitting in and turn it into a pure female. And you can see that dramatically in the external morphology of the plant.

This same process goes on in insects and in crustaceans, when they are infected by a cytoplasmic bacterium called Wolbachia. It is in Wolbachia's interests only to occur in females, because they can only get into the next generation in eggs; they can't get into the next generation in sperm. And Wolbachia will feminize the organisms that it's in, and in some cases Wolbachia will kill the males, so that the male offspring don't develop. So these are cases where the conflict of interest is arising because there's selection going on at two different levels; on the whole organism, and then within the cell, on the cytoplasmic organelles.

So if we look into our own genome--and I'm going to spend the last twenty minutes of this lecture looking into our own genome--we see a very interesting thing. If there's potential, through either hierarchical selection or asymmetry of information transmission, to generate evolutionary conflict, then we see that we're not even in principle the consistent wholes that you might think we are. And a very famous guy said that, "Perhaps this is some comfort when we face agonizing decisions, when we cannot make sense of the decisions we do make, and when the bitterness of a civil war seems to be breaking out in our inmost heart."

And that was Bill Hamilton, the guy that came up with kin selection, and wrote a lot on evolutionary conflict. And it's fairly poetic. And Bill actually liked haiku a lot. He particularly liked Basho's famous Narrow Road to the Deep North, which is--Basho was one of the greatest Japanese haiku writers, and was a favorite of Bill's.

So there are interesting implications of what we're talking about today, and the outline basically is going to be how you can generate genomic conflict out of hierarchical selection. I'm going to make a strong point that the opportunities for conflict are much greater in sexual than in asexual species. Then I'll mention that the uniparental transmission of cytoplasmic genomes is probably a method of conflict resolution. Then I'll go on to talk about genomic imprinting and parent-offspring conflict in mammals. And then that outline there represents what is well-established and reliable. When I go off this outline, at the end, I'm going into the speculation.

So conflict can arise in two situations. One is the Russian doll situation, the babushka situation; multilevel hierarchical selection. That is when one selection process is contained inside another selection process; and here you should think of things like meiotic drive and cancer.

The other situation is where the transmission is asymmetrical, so that the different genetic elements in the system do not all follow the same transmission pathways. The cytoplasmic organelles are the classic example. They can only go through the female line, they can't go through the male line. The nuclear genes are going equally through both male and female inheritance. So there's a large and striking difference in the way the cytoplasmic organelles are inherited.

So when we think about two-level selection, there are really two things that can be going on. For example, here we have two genetic entities contained inside a larger thing. Okay? If A has a replication advantage, at the lower level, then it can just build up more copies of itself, and then when this larger thing divides and reproduces, it will end up in more copies, because at this stage it was reproducing faster.

Think petite mutation in yeast. A petite mutation in yeast is a mitochondrial mutation, and basically what the petite mutation does is it cuts out a chunk of the DNA in the mitochondrial genome, so that the mitochondrial genome can be replicated faster. Now, of course, if you cut out a bunch of the mitochondrial genome, the mitochondria aren't doing their job of being a good energy factory so well, for that cell that they're living in. So they're gaining an individual advantage from mitochondria, but they're damaging the interests of the cell that contains them. And what happens is that the ones that cut out the DNA, that can replicate faster, do build up a replication advantage at the lower level.

The other possibility in two-level selection is that there's a segregation advantage. There are just as many copies made of each type, at the lower level, but in the process of then forming say the gametes--so in any replication process, either mitotic or meiotic, if there's a segregation advantage, one of them is going to get into more copies. So it takes the same number, and then just in the process of making the new cells it gains an advantage. And think here meiotic drive. Okay, so that's the paradigmatic example for segregation advantage.

So in the petite mutation in yeast, what's going on basically is that there's a deletion in the mitochondrial genome. That allows the shorter genome to be replicated faster. It builds up a big population in the cell. However, there's a disadvantage at the higher level, and that is defective metabolism. The result is that the cell lineage goes extinct. And so people who work on yeast in the lab--if you just take a big population of yeast and you played it out in generation after generation, these petite mutations keep popping up, and they spread, and then they go extinct. They have a lower level replication advantage, but they have a high cost for the cells that contain them, and they disappear. It's almost exactly analogous to cancer.

In an asexual lineage, the only kind of conflict that is in principle possible is one selection process contained inside another one, and the conflict would occur if the lower level response differs from the higher level response; so if what's good at the lower level is bad at the higher level. Petite mutation is a good example.

There's no horizontal transmission there, because there's no sexual reproduction going on. So two independent lineages are not coming into contact with each other and mixing; they're staying separate through generations. And so there isn't any way for the lower level response to escape the fate of the upper level response. So if there's a significant conflict, a significant cost, the lineages will die out. So this is something that actually can drive asexual extinction.

In a sexual lineage, sex is creating genetic variation within the nuclear genome. It has the potential to create genetic variation in cytoplasmic genomes, and it creates opportunities for non-chromosomal genetic elements to change hosts; particularly interestingly in bacteria it does this. So some kinds of mechanisms, that are going on during sex, formally resemble pathogen transmission; the transmission of--when you cough up a virus and it gets into your roommate, basically a genome is moving horizontally from your body into another body, and reproducing there, and during sex there are opportunities for this kind of thing to go on, from one bacterium to another, and certainly in organisms like us, from one organism to another.

So one cost of sex might be the potential it creates for inter-genomic conflict. I'm not talking here directly about sexually transmitted diseases like gonorrhea or syphilis. I'm talking about the possibility that genetic elements infect the genomes of other cells.

So, for example, there could be a conflict between bacterial plasmids and chromosomes. A little background on bacterial genetics. Bacteria usually contain plasmids, and these things are small circular genetic elements and they live in the bacterial cytoplasm. So you can think of them as genetic parasites. The rest of the bacterial genome is a large single circular chromosome which is attached to the cell wall of the bacterium. So think of the bacterium as a balloon that has a circular rubber band attached to the cell wall, but then out floating in the balloon are these much smaller plasmids that contain DNA and do particular things.

The plasmids often are the elements in the bacterium that have genes for antibiotic resistance, and they can be advantageous when antibiotics are present. There are other plasmids that will addict their host cells, the bacterial cells, to their presence by making a poison antidote system. Okay? So basically what they're doing is that they're protecting their own cells and they are producing chemicals that destroy cells that don't contain the plasmids.

And this is a general principle. If you make a long distance poison and a short distance antidote, you protect the environment that you're in and you destroy the competition. So any bacterial cell that doesn't inherit the antidote, via a plasmid, but gets the poison, will die. So that changes selection dynamics, at a higher level, and this plasmid will spread through the population.

Very similar to that, in some sense, is segregation distortion. There is a gene that was first found in mice and--this is important--it's just an arbitrary accident of developmental biology that this segregation distorter happens to also result in mice with short tails. Okay? That's just an accident of pleiotropy. This gene has effects on both segregation distortion and on tail length, in mice. So you can think of the fact that it's affecting the tail as just a marker; it's just kind of like having a reporter gene in there. And we'll simplify the situation and just consider two alleles. There's a t and a normal allele that we call +. Okay, so these are the two versions of this gene that are sitting at the same place in the chromosome. If you have tt homozygotes, they're lethal or sterile. So if that were the only thing that were going on, you'd never see this thing; it would die out real quick.

But if you have a mouse that is heterozygous with t and +, they produce--they're fine, they live just fine--and they produce 90 to 100% t-bearing sperm. This is again done with a long distance poison and a short distance antidote system. So sitting there in the testes of the mouse is a cell that is making sperm, and some of them have the t and some of them don't, and the sperm that have the t in them are making poison, which is going over and killing the ones that don't have the t. They're sitting spatially right next to them in the testes, and the sperm that have the t are also making an antidote, but it's only effective inside their own sperm. So basically what's going on is that t's just wiping out the competition, inside the testes, and that ends up producing 90 to 100% t-bearing sperm.

So you have hierarchical selection at the level of the gamete. You have got selection for t and against t, up at the level of the diploid individuals, because up there the tt homozygotes are lethal or sterile. So a 50:50 sex ratio--ignore that, this is irrelevant right here; this is for, this sentence snuck in here for a different kind of gene action. So this sentence is--and I regret that, I should've edited that out.

If the tt homozygotes didn't die, but they suffered a sufficiently small, sub-lethal fitness reduction--so if this part here were not true--then t would spread, and eventually if t spreads all the way through the population, everybody's got the antidote, and you don't have any segregation distortion anymore, and everything goes back to normal. If that is the case, once t takes over the population, there's no more segregation distortion.

This introduces the interesting possibility that most species may have had a history of segregation distortion and we just don't notice it anymore, because they've gone to fixation. In fact, we don't have any easy method of detecting that. We may see the traces of that, written in the history of things like the fairness of meiosis, but we can't go out right now and easily find genetic or biochemical evidence that we have fossil segregation distorters sitting in our own genome. It seems likely that we do, but we don't know.

Now what about conflicts between the nucleus and the cytoplasm? Well any cytoplasmic genome, that's replicating faster, gets a segregation advantage, because there isn't any meiotic mechanism that assures fair segregation of organelles. The chromosomes are controlled by the spindle apparatus. They line up at the plate, at the middle of the cell. They make two copies. The spindle grabs one copy and pulls it one way, and the other copy and pulls it the other. Okay? So that's really fair, that's exactly 50:50.

The organelles are out there floating around. They're not attached to a spindle when the cell divides, and so basically if they can just make more copies of themselves, they'll have a better chance of getting into the dividing cells.

If you had biparental inheritance of cytoplasmic genomes, that would mean that in the same cytoplasm you would have genetically different, unrelated mitochondria; genetically different, unrelated chloroplasts. And the consequence of that would be conflict, and that would be expressed as an organelle cancer. If you only get the cytoplasmic genome from one parent, then they'll very likely all be the same genotype--any kind of process like this going on in the past would have assured that there would only be one left standing, in that parent--and therefore they're not in conflict with each other.

So, in fact, you all only contain mitochondria from your mothers. It's extremely rare that a human will ever have a mitochondrion from a father. It does happen, but it's a one in a billion chance. Okay?

So those are some of the well established cell level scenarios in which conflict plays out. And before I go into the reproductive problems in humans that result from conflict, I'd just like to emphasize that this vision of evolution doesn't sound like the beautifully adapted world where all is for the best, in the best of all possible worlds. This is a vision of evolution in which there is continual conflict, and in some cases it's never resolved, which means that in some cases both sides are paying a continual price. So that's quite a different way of looking at the world. And if you're trying to derive simplistic take-home points, from the evolutionary view of the human condition, one of them would be, as you'll see in a few minutes, that there are probably long-term conflicts that are never resolved.

So reproductive problems. In mammals there are conflicts between mother and fetus over how much the mother should invest in the fetus. The symptoms of that are pre-eclampsia and diabetes. There are conflicts between mother and father over maternal provisioning, and those are related to genetic imprinting of growth genes, and there are disturbances and a tug-of-war balance produced by evolutionary conflict in genes that are expressed in the infant brain, and those are thought to deal with mental illness. This is where the line is between well-established science and speculation.

So the arenas in which these things play out are in the placenta and uterus, and in the developing brain. And this is the sequence of ideas; so I'll give you a little intellectual history. 1961, '62, Bill Hamilton has the idea of kin selection, the idea that we can- a gene can increase its fitness, either by its actions on my own body, or by influencing the actions that I take to improve the reproductive success of relatives in which that gene also probably exists. Then Bob Trivers developed Bill's idea into parent-offspring conflict. And the idea of parent-offspring conflict--which I'll state a couple of times to make clear--is this.

A mother is 50% related to all of her offspring. She is interested in making sure that each of them has an equal chance therefore to have grandchildren. Now switch your point of view to one of the offspring. It's 100% related to itself; it's 50% related to a full sib; and it's only 25% related to a half-sib. So from the point of view of a gene which is sitting in our focal offspring, it wants to titrate its mother's investment away from its potential future siblings and into itself, until the probability of grandchildren, through itself, exactly matches the probability of grandchildren through the others multiplied by degree of relationship. Okay? It will have a full-sib if the species is monogamous, and it has a probability of half-sibs if the species is polygamous; if the mother mates with multiple males. So that was Bob's insight.

Bill got the Crawford Prize, which is the Nobel Prize in Evolutionary Biology, for kin selection, and Bob got it for parent-offspring conflict. So those are prizes that are worth oh six or seven-hundred-thousand dollars, and like Nobel Prizes, they are awarded in Sweden, in Stockholm. And so these were seen as big, important ideas.

Now David Haig then picked up on Bob's idea, and he said, "Well, there's not only conflict between parent and offspring." And that conflict, by the way, is also realized through imprinted genes in pregnancy. There is conflict between the mother and the father over how much the mother should give to the baby, and the baby take from the mother.

So if the father can put into the baby a gene that then extracts more from that mother than the mother wants to give, the father can gain, to a certain point, an advantage. This isn't an absolute thing, it's just saying that there is a range of investment where it is not advantageous for the mother to give more to the baby, but it is advantageous for the father to get the mother to give more to his baby. Okay? And this is mediated by genomic imprinting.

The final step in this little bit of intellectual history is Bernie Crespi and Chris Badcock, who came up with the idea that this conflict that David Haig identified, which is going on during pregnancy and is probably mediated mostly by genes that are having interaction in the fetus and in the placenta, extends into early life during the period of suckling, before the child is weaned, and the conflict is then expressed in genes that are in the brain of the infant, and when their tug of war, which is in evolutionary equilibrium, is disrupted, Crespi and Badcock think that you get mental disease.

So this is Bill. This is taken on the Amazon. Bill died in the year 2000, after trying to find the source of AIDS in the Congo. He had gone to the North Kivu to see whether or not he could find chimpanzees whose DNA might match DNA in polio vaccine from the late 1950s. There was a hypothesis at that time that that's how HIV got into humans, through polio vaccine. It turned out to be wrong, and Bill died just after that trip.

This is Bob Trivers, recently. He's a prof now at Rutgers, and Bob had the parent-offspring conflict hypothesis, as a grad student at Harvard in 1969, '70, '71; about there. This is David Haig, currently a professor at Harvard, and the guy who came up with the observation that there's a very intriguing connection between the imprinting, the differential imprinting of genes in the male and the female germ lines, and the control of growth by the embryo. And this is Bernie Crespi and Chris Badcock. So Bernie's at Simon Fraser in Vancouver, British Columbia, and Chris is in London, at the London School of Economics. So these are the guys who had these ideas about evolutionary conflict, expressed in humans.

There was a news item on the local television station the other night that Jacob Lykke's research on preeclampsia had just been published, and it was sort of playing up the idea that research from Yale Medical School reveals important pregnancy complication consequences; women who have preeclampsia have worse health later in life. That basically is a continuation of this idea. The OB/GYN Department at Yale Medical School has picked up on this stuff.

Okay, so let's run through the logic. The conflict between mother and fetus over maternal provisioning is basically this: the fetus is selected to extract more from the mother than the mother is selected provide. It's 100% related to itself. She's 50% related to each of her offspring. It wants to take more from her. She wants to hold some back, so she can give the same amount to future offspring. The way that it will do this is by using tissue in the placenta to secrete hormones into the mother to manipulate her metabolism.

It also does it, by the way, morphologically. It is fetal tissue in the placenta that invades maternal tissue, aggressively, and establishes tighter and tighter connections with the maternal blood circulation. So if you look at the origin of the cells in the placenta, there's a morphological story of conflict written there as well.

So the symptoms experienced by the mother are high maternal blood pressure and pregnancy-related diabetes, and this will happen particularly when this gets a bit out of balance. So if you're a baby, sitting there in the womb, and you want to get more out of mommy, you can do two things. You can pump up her blood pressure so it'll force more nutrient through the placental barrier, and you can play with her metabolism so there's more sugar in her blood. Too much of that and the mother gets pretty sick.

This is the arena in which it occurs. This is the fetal portion of the placenta here; you can see the invasive blood vessels going in, over here. This is the maternal portion out here, and this is where the exchange of nutrients is mediated.

So the evolutionary logic behind this is that--if we now look at the mother-father conflict--the father isn't going to be related to the mother's later offspring, if they have other fathers. And, by the way, I'm now going to make a series of statements that sound like humans are engaged in absolutely outrageous moral practices.

None of this logic necessarily is going on in current evolution, in our current human population, because you can demonstrate these effects in mice, and we shared ancestors with mice about sixty million years ago. Okay? So a lot of the machinery that's being detected in mice and in sheep and in humans, that is shared, could very well have had to do with the polygamy, or lack of monogamy, in ancestral mammals a long time ago. Or it could still be going on.

Now there is an asymmetry in the male and female reproductive possibilities. The father's reproductive success depends on his successful matings. The mother's reproductive success depends on the number of offspring she personally can bear. And if you state that brutally, he can have several children and other females, while she's dealing with this one.

So here is a Mormon polygamist. He has two wives. This brother is 50% related to this sister, and 25% related to this brother. Okay? So bear that scenario in mind. That is the sort of thing which is driving the selection pattern. Rare today in humans; possibly much commoner in the past.

What does this have to do with imprinting, and what is imprinting anyway? Imprinting is a process of methylating genes, and if you imprint a gene, you turn it off; it will not be transcribed if it's methylated. Imprinting is used in a number of contexts. It's an epigenetic mechanism that's used in development to control cell fate.

But the kind of imprinting that we're talking about today is a special kind. It's differential imprinting by sex, and it's not happening during the development of the body in order to decide whether a cell becomes a liver cell or a brain cell, it's happening in the germ line of the parents, just before the gametes are produced. And the point is that the father is imprinting certain sets of genes and turning them off, and the mother is imprinting other sets of genes and turning them off. Okay?

These genes that are imprinted in the germ line are not expressed in the fetus, and they are then reprogrammed in the germ line of the adult. The adult could be either male or female. Right? So when it makes its gametes, in the next generation, it doesn't make them with programming the imprinting pattern it had when it was a baby, it makes them with the imprinting pattern that is appropriate to its sex.

What's going on is this: The father is turning off genes that down-regulate growth in the embryo. The mother is turning off genes that up-regulate growth, and so basically--it's kind of double negative, because the father's turning off stuff that acts in the mother's interests, and the mother's turning off stuff that acts in the father's interests. But the upshot of that is that the father is trying to program the embryo to extract more than the mother is prepared to give, and the mother is resisting.

You can only see this going on when you disturb the equilibrium. You can disturb the equilibrium in a number of ways. You can do it by genetically transforming mice, and the gene that you choose to disturb the equilibrium is the gene that does the imprinting. Okay? So you mutate that gene, or you delete it, and then you observe the outcome. And the effect is roughly plus or minus 10% in birth weight.

So if the father's genes are--if the mother's genes are not doing their job, so that only the father's interests are expressed, the embryo is about 10% heavier; and if it's the other way around, the embryo is about 10% lighter. This scenario is also supported by the fact that if you look at all of the genes in the body, there are only about 100 or 200 that are imprinted. There are very few that are imprinted differently in the mother and in the father, and the ones that are imprinted differently in the mother and the father mostly have to do with the control of fetal growth.

So it's a very special set of genes, and they are clearly associated with the specific function of fetal growth, rather with the millions of other things that genes do in the body. So up to here, you can do nice manipulation experiments on mice, and these have also been done in things like sheep, and the scenario stands.

Now, going into speculation. The primary site is the placenta, where there are small deviations that can benefit child or mother. Large deviations are costly to both. So the normal situation might be that there would be a small deviation. You only get a big deviation and disease when there's a real disruption of the imprinting patterns and they get out of balance; so that if you're thinking of a tug of war, one side falls over. Okay?

Where are the rest of these genes? They're in the brain. These are the sex differentially imprinted genes, the ones that are imprinted differently in mother and father, and they're not controlling fetal growth or being expressed in the placenta, but they are being expressed in the brain.

And this is now Crespi and Badcock's idea. A deviation toward paternal gene expression should result in a relatively selfish offspring. So it should be trying to take more from the mother, and it should be doing it now through infant behavior rather than through fetal physiology. And a deviation towards maternal gene expression should result in an easy offspring that would be letting the mother relax and store up nutrition for the next baby.

So we can't do experiments on humans to--like we can say with knockout genes in mice. So what Crespi and Badcock have done is they've looked at neurogenomic syndromes, single gene effects, and idiopathic psychiatric conditions, to see what happens when this tug of war in the brain is disturbed.

Well probably the most revealing early observation--this was actually picked up by David Haig, before Bernie got into this--is that there are imprinted genes on chromosome 15, that are expressed in the brain, and if the maternal copy is deleted or modified, you get one syndrome, and if the paternal copy of this same gene is deleted, you get another syndrome.

So Angelman syndrome is that the maternal copy is deleted. The paternal copy is only imprinted in the brain; it's not imprinted in other parts of the body, it's very specifically imprinted in this tissue. And Angelman children are happy, retarded and uncoordinated. The same gene, but with paternal copy deleted, maternal copy imprinted, you get, after the age of two, you get uncontrolled eating, hypogonadism, delayed puberty, and a completely different syndrome. Okay?

So the Angleman types, with father's interests over-expressed, have prolonged suckling, frequent crying, hyperactive, sleepless; they're difficult children and they have high rates of autism. And the Prader-Willi children, with maternal interest over-expressed, don't feed very well, they cry weakly, they're inactive or sleepy, and they have high rates of psychosis; some kinds of psychosis are called schizophrenia.

So what Crespi and Badcock proposed is that if there's an imbalance during fetal development, in the brain, towards paternally expressed imprinted genes, you get higher birth weight, a larger brain, faster growth, a cost to the mother. The costs to the mother are coming from selfish, egocentric cognition and behavior, and both mother and child are bearing costs from any of the negative aspects of autistic spectrum. Okay?

If the mother's interests are over-expressed, then during fetal development you get a smaller birth weight; you get a smaller brain, less lateralized brain; slower growth. The benefits to the mother is that the child is easier to take care of. There's a cost to the offspring, it has schizophrenic behavior. And there's also eventually a cost to the mother, from the schizophrenic behavior.

So the people who have featured here--this is just an autistic child sleeping on his hands. That's Sylvia Plath, the poet who committed suicide. And if we look at the correlative evidence from idiopathic schizophrenia and idiopathic autism--so I think you know who those guys are--what you see is that associated with schizophrenia of unknown cause--that's what idiopathic means--is low birth weight; slow growth; small head/brain size; better verbal; dyslexia; some overlap in the genes with bipolar disorder; major depression.

And if you look at autism, you see that they have higher average birth weight; so not always higher, but not lower. They have faster body growth. They tend to have large heads. They are called hyperlexic. They have better visual, spatial and verbal skills, and you can get an idiot savant syndrome out of that.

So there is correlative evidence. This is not experimental, but it is possible to go out there in the literature and to pull together a lot of studies and say, "Hey, it looks like there are correlations with what you might expect if this was an over-expression of maternal interest and this was an over-expression of paternal interest in the infant brain."

Now if this connection between evolutionary conflicts of interests and mental disease is ever actually established, it's going to be one of the most remarkable connections that I know of. It was completely unexpected. Nobody ever thought that an alternative explanation for autism and schizophrenia would ever come out of kin selection and parent-offspring conflict. Okay? Certainly that was completely unsuspected in the '60s, '70s, '80s and '90s.

So I'd now like to pause and just remind you that everything that I've told you about a potential connection between evolutionary conflicts of interest and mental disease is speculative. It is actually at the moment an object of rather intense research. But the annals of research journals are littered with the corpses of beautiful ideas that were killed by facts, and that could very well happen to this one. We have to be patient and just see what happens. But I hope that I've been able to communicate to you that there is a role in science for bold speculation, and that it actually makes the whole process extremely interesting.

Now I'd like to do something that occurred to me after I had lunch with Bill Feldman yesterday. Bill's taking this course because he's a political scientist, and he's interested in what evolution has to say about politics. So I want to give you some take-home messages about conflict resolution that come out of the study of genomic conflict.

If you want to get rid of a conflict, make the interest of the competing elements symmetric. You can do this in a host pathogen relationship by shifting the transmission from horizontal to vertical. That will reduce their virulence. Because, if you think about it, then the pathogen can only get into the next generation if its host survives. If it's a vertically transmitted parasite, that means it's transmitted from parent to offspring. So the parent has to survive, to have a baby, so that the pathogen can make it. So it's not in the pathogen's interests to kill the parent. A horizontally transmitted pathogen, on the other hand, can have actually quite a high level of virulence; and that is where all the major diseases are, they're all horizontally transmitted pathogens.

And if you think about things like Wolbachia--remember, I was telling you about this bacteria that feminizes its hosts, so that it would always be occurring in the body of a female. Well there's some crustacea that have figured out how to solve this problem. They take the Wolbachia and they chop out its sex determining gene, and they implant the Wolbachia sex determining gene on one of their chromosomes, et voilà, there is no conflict anymore because now the whole business is vertically transmitted.

So they just took the offending element, that one offending element out of the bacterium, and they stuck it into their nuclear genome, and they created a new sex chromosome for the crustacean. So they also got--they made the interest symmetric. Both genetic elements then had the same vertical transmission route.

Another way that you can resolve conflict is this. You can suppress the meiotic drive. So you can punish the offenders. And the evidence for that basically--that there's been a history of suppression--is the fairness of meiosis. You can also homogenize the reproductive success of competing elements within a group, at the human level. This can be done with monogamy. Anything that will make individual success depends on group success.

So I'm going to give you a couple of taglines to remember this, a couple of mnemonics. A rising tide lifts all boats, and we're all in the same boat. So if you're a gene, you should think that anything that you can do to improve the reproductive success of the organism that you're sitting in, is probably the only way you'll improve your own, and the fact that it's improving everybody else's reproductive success, in that same genome, is actually irrelevant to you. You're not competing with them; in fact, you're all cooperating, because if you're all in the same boat, and you're all pulling together, that's the only way to get into the next generation.

And, of course, there is this anecdote--this isn't--we don't actually know if this is a direct quote or not. But it is said that as the Declaration of Independence was signed on July 4th in Philadelphia in 1776, Benjamin Franklin turned to the people who were signing it and said, "Gentleman, we must indeed all hang together or assuredly we shall all hang separately." So these are just mnemonics. These are ways of remembering the principle that the way to suppress conflict is to generate a situation in which everyone is dependent upon partners for success.

So the take-home for the lecture basically is this. You should think of organisms as a hierarchy of replication levels, and natural selection can occur simultaneously at all of the levels in the organism. This is especially important with cytoplasmic organelles and with meiotic drive.

Replicating units that only occur in a few copies, and whose replication and segregation are strictly controlled--things like cell nuclei and chromosomal genes--do not easily cause genomic conflict. But if those units occur in many copies, and if their replication and segregation is not strictly controlled--those are things like cytoplasmic genetic elements--they more easily cause genomic conflict. Conflict is much more easily evolved and experienced in sexual organisms than in asexual organisms. Okay.

[end of transcript]