Mutation rates and paternal age


, , , , , ,

I briefly want to talk about a newly minted shiny article published in one of the scientific ‘glossy’ journals, those high-profile journals that lead to the bulk of the science-news coverage. This one was published in Nature a week or so ago.

I wanted to wait a bit before writing about it, and now that we are nearing the end of the typical article news cycle (1-2 weeks) it is time. Here is the main conclusion from the article: older fathers pass-on more new mutations to their children than younger ones. The most important background fact is that mutation is the stuff of evolution. It is the raw change that allows all organisms on earth to adapt. Biologists generally hypothesize that mutation rates are constant, meaning that DNA changes accumulate at a certain rate as organisms age. In a recent work Augustine Kong potentially challenged that idea (see main article figure below).

English: DNA replication or DNA synthesis is t...

Some mutations are linked to disease, a child with more mutations is at higher risk, just by laws of probability, of getting the ‘disease’ mutation. Importance of this data to the realm of human disease is obvious. The substantial media coverage following Kong’s publication almost entirely focused on the disease aspect. Here are some headlines from the usual suspects: “Older fathers linked to Kids’s Autism and Schizophrenia risk” says Time Magazine, “Older dads may raise risk for autism in kids” adds FOX, “Father’s age is linked to risk of autism and schizophrenia” finishes the New York Times while omitting the kids aspect in their title all together.

Here is what the headline, in my opinion, should of read “Mutation rates are not constant, new Iceland population study suggests.” It is not an especially catchy title and I see that. Disease is bad for people, good for biologists. Biologists sell their work and build careers by putting words like, disease, autism, dawns, and MS into the titles of their papers and grants. Augustine Kong is not first to insight  media frenzy with ‘disease’, and that’s OK because journalists got to do their job and no story sells better than a story that everyone is afraid to hear.

I read the journal article and swam the sea of biased media coverage waiting for the bile and rage to loosen its grip. Then I wrote this, and tried to mention the real interesting bit, the implications of this work on how biologists estimate divergence. Mutations are assumed to accumulate in a clocklike way, same rate over time for each type of organism. This allows comparisons between organisms, like humans and monkeys and squid and bacteria. Because of this constant ‘mutation clock’ biologists are able to say that genetically humans and monkeys are more similar that humans and tomatoes. Because of this ‘mutation clock’ biologists can estimate how long it takes for organisms to diverge and become different enough to be considered different species.

What Augustine Kong and friends showed is that the mutation rate is not constant, and they were as far as I know the first ones to actually calculate the rate of mutation increase with age.

English: zebrafish histology atlas; testis; sp...

With age fathers ‘give’ more new mutations to their children. Mechanistically, from the cell biology point of view, this implies that as males age they incorporate more genetic mistakes during sperm production. This is also extremely interesting. In males, the machinery that proofreads DNA replication during sperm formation may degrade with age. This last one is a crazy idea and there is very little substantial proof behind it, but it points to an interesting question of how these new mutations appear.

I want to end with this final point, it may be subtle and it is definitely intuitive. Demography matters. Kong’s work did move biology forward, and strengthened the link between the important ideas of genetics and demography. It is important when we mate and how many offspring we produce and how well we feed the first and the last ones.

Contributor Artur Romanchuk is a fifth-year graduate student at UNC Chapel Hill studying with Christina Burch and Corbin Jones. Artur primarily studies how bacteria pass genes to one another (“lateral transmission”) and how these new genes lead to the ability to infect new hosts. He is also an author and cartoonist: check out his other works at ingradients and WHandCats. His first daughter was born when he was in his mid-twenties, and should be relatively mutation-free.

The cost of reproduction in birds


, , , , , ,

The concept of trade-off is paradigmatic in life-history theory. an organism can only acquire a finite amount of energy in its lifetime, so it must “choose” how to allocate that energy to growth and survival or reproduction. Reproduction is assumed to be costly so that individuals who spend more on reproduction, for example by laying more eggs, will not survive as well. We suppose that over evolutionary time, natural selection will act on genetic variation for these allocation decisions, so that the sequence of decisions over an individual’s lifetime will represent an optimal allocation of resources.

Unfortunately this intuitively appealing idea has been very hard to find in nature. In fact, many studies have come up with positive correlations: animals that reproduce more tend to survive better. A recent study by Eduardo Santos and S. Nakagawa found that this trade-off was almost impossible to detect in most studies, or non-existent altogether. In a meta-analysis of brood supplementation studies (researchers added eggs to the nests of breeding birds), they found little impact on survival. Their result held across all the major taxonomic groups of birds, the biggest division being between passerines (songbirds, crows, flycatchers, etc) and non-passerines (ducks, loons, parrots, woodpeckers). Regardless of overall “lifestyle” the birds tested in most studies were able to withstand the hypothesized survival cost of additional eggs dumped on them by researchers.

Bird - Seagull enjoying the sunset

Why would this be the case? As always there is the possibility that the studies were poorly designed, or that brood supplementation is not a good way to test for a trade-off. Particularly, brood supplementation only taxes the parents of their ability to defend and feed offspring; it does nothing to the energy that females put into egg production. The other possibility is that adult birds just don’t put that much effort into reproduction in the first place. Perhaps survival is far more important. The trade-off is still there, but it’s just not important for most birds.

The hypothesis that life is just not as Malthusian as we have often supposed in evolutionary biology intrigues me greatly. If evolution acted in the “well-oiled machine” manner that many laypeople and professional scientists find appealing, then we’d expect selection to push annual reproduction right up to the level allowed by the trade-off. What studies have found is birds putting minimal effort into reproduction, parenting or anything that affects their survival. This means that selection is a lot weaker than we expect: this gives genetic drift a lot more room to account for polymorphism. It also makes sexual selection more plausible: if most species have fairly conservative lifestyles and selection for survival is not that strong, then males (or females) can afford costly ornaments.

An unrelated study also appeared this week that is getting a lot of press: researchers in Iceland found a strong relationship between the age of fathers and mutations passed to their offspring. This is the first study to quantify the per-year effect of paternal age on offspring mutations in humans, so it’s a pretty big deal. I will talk more about this in a future posting since it’s related to my dissertation research, but in the meantime, go read the article and enjoy the flurry of debate surrounding it.

E. S. A. Santos, S. Nakagawa (2012). The costs of parental care: a meta-analysis of the trade-off between parental effort and survival in birds Journal of Evolutionary Biology, 25, 1911-1917 DOI: 10.1111/j.1420-9101.2012.02569.x

Drift and selection: the epic battle?


, , , , ,

I learned a lot at last week’s Nescent Academy on Quantitative Genetics. I saw a lot of material that I wouldn’t have seen in other forums, like the Ornstein-Uhlenbeck models of genetic change under microevolution and macroevolution. During the last half of the week, which focused on macroevolution, I confirmed my impression that when talking about evolutionary history, genetic drift is really the name of the game. When population geneticists talk about the history of particular genes (for example, a gene implicated in a human disease), they rarely speak about selection. There was also a lot of good information on research findings from natural populations.

There were three points in particular that struck me:

  1. When looking at natural populations, there is abundant genetic variation in almost any trait
  2. Selection is generally weak and generally stabilizing selection
  3. Stochastic processes, such as genetic drift, can account for a lot of diversification seen in nature

These findings were interesting to me because I study selection, usually using deterministic models and because I’ve seen the perspectives of other researchers about the relative roles of selection and drift. I have tended to assume, along with many other researchers, that most diversification is due to selection, and that for any “real” differences to matter over evolutionary time that selection must be involved. Why is this?

Selection and adaptation are appealing concepts and they are simple to understand. Darwin’s (three or four) postulates give us all we need to understand how adaptation comes about. Adaptation is a really nice idea: things become more efficient, better, over time. Not only is that aspect appealing, but it’s easy to understand how it could happen: selection eliminates the less efficient, and promotes the more efficient. All you need after that is inheritance. This is so easy that most people get it the first time. Leaving aside the appeal of this from the social perspective (read the first chapter of The Dialectical Biologist), it’s just easy. I teach evolution and ecology to undergraduates and most of them come in getting the basic idea of selection. It’s not hard.

Take genetic drift on the other hand. If you’re like most biologists who’ve tried to teach about genetic drift, you know that genetic drift is the opposite of selection from a teaching perspective. Genetic drift, like selection, removes variation from populations. Only mutation can bring it in. Under genetic drift alleles just disappear: by random chance they fail to make it into the next generation. This only happens in finite populations, that is every single real population. How? Think of it this way: you know that if you flip a coin a thousand times you will get close to 500 heads and the rest will be tails. Do this instead: flip a coin ten times ten times and record the number of heads you get for each ten coin flips. You could then make a graph depicting the number of times you get five heads, six heads and so on. You should get a nice looking histogram: close your eyes and put your finger on a spot on the graph. Your real population is that spot. It could be the one where you got zero heads.

There are two important things about genetic drift: one is how it leads to diversification, and the other is how it accounts for polymorphism. Drift leads to variation between populations because when populations are separated they randomly undergo drift, possibly with different end results: if there are two alleles A and B at a locus under drift, one population could lose allele B and the other could lose allele A. Repeat that over many loci and your get very different looking animals that don’t recognize each other when they get a chance to make babies. The second property is that when you observe polymorphism (genetic variation), it is probably due to drift. Drift over time removes genetic variation from a population, but before that happens the frequency of the allele in question bounces all over the place by random chance. The time window over which that happens is incredibly large, much longer than that for selection. Therefore the large amounts of genetic variation in natural populations are probably due to weak selection, strong drift and lots of mutation across the genome.

Here’s my explanation of the above findings: selection is always happening, but is generally weak. Selection is weak both because drift is happening at the same time, and because life is just not as hard as Darwin and Malthus had in mind. Selection in nature is usually stabilizing selection, meaning that there is some intermediate value that is favored most, and extreme values are selected against: the typical example is birth weight in industrialized societies. Small babies are prone to infection and pulmonary dysfunction and large babies are at greater risk for perinatal complications. However, in most cases, it appears that deviations from the optima that we can detect in nature are not heavily penalized. Especially in large-bodied, iteroparous organisms like birds, ungulates and primates, life just doesn’t seem that hard. This means drift has plenty to work with. Most of the organic diversity we see is probably due to drift randomly sending populations closer to new optima where stabilizing selection takes over again. This is basically Sewall Wright‘s shifting balance theory.

Wright in 1954

Wright in 1954 (Photo credit: Wikipedia)

This positively demonic process could account for most of the organic diversity we see out there, but it is not an appealing idea. I think most intellectuals go through a phase where they attribute everything to randomness — and I’m not suggesting we all get on that bus — but there’s also a Conspiracy to remove slack from the world. People generally don’t like the idea of random forces to explain things. Especially since a lot of biologists, including myself, don’t understand genetic drift (how could you?) it’s really hard to get behind the idea. However, especially when analyzing real populations, such as the evolutionary history of humans, and testing ideas about sexual selection, we have to consider the role of drift. Much of the persistent, between population variation we see that looks adaptive could be due to genetic drift.

What is quantitative genetics?


, , , ,

Next week I’ll be attending a Nescent Academy “master class” on quantitative genetics, led by two teachers who have made major advances in this field. They will be assisted by other teachers who have done important work in the field. I’m really glad to have the opportunity so close to home: it’s always nice when the best students and teachers in a field come to my home town to lead a workshop like this 😉

This is a good opportunity to talk about quantitative genetics. What is it? As with many scientific topics, it means different things to different people, depending on your main research question. I do theory, but other researchers are primarily concerned with analyzing data, and conducting experiments. Quantitative genetics is a helpful set of ideas in all three of these areas, so I will try to explain my understanding of all of them.

First of all what is genetics? Genetics is the science of inheritance. The basic question of genetics is how do parents pass their traits to their offspring? We know the basics are that parents pass “particles of inheritance” to their offspring (genes) that the offspring express, thus creating likeness between offspring and parents. This theory works for traits like eye color and hair color, but what about for something like height? Height is a different kind of trait altogether, and people have known for a long time that offspring do not typically display the height of one parent or another, the way they do with eye color.

This disparity was actually the source of a vitriolic debate between “Mendelians” and “biometricians” at the turn of the twentieth century. The biometricians had been studying traits like human height in humans and animals for many decades, and then several botanists rediscovered the work of Mendel and started testing it out on plants and fruit flies. Mendelians caricaturishly believed that mutation was the only necessary evolutionary mechanism, and that Mendelian inheritance was the only mode of inheritance possible. They couldn’t explain the patterns of inheritance in traits like height, however.

The solution was a “fudge” or a “hack” by R.A. Fisher: what if height was controlled not by a single gene, but by many throughout the genome? Offspring will inherit some of these from mom, some from dad and the outcome should be a mix of contributions to the trait from each parent. Height, after all, or any trait, is a human construction, imposed by researchers onto an organism: why should there be just one gene that controls something arbitrarily decided by someone with a measuring tape? The outcome, Fisher showed, of Mendelian inheritance of a huge number of genes contributing small effects would be just the patterns observed by the biometricians. Keep in mind that in the first half of the twentieth century, people did not know what the actual genetic material was, or least of all how it worked.

Quantitative genetics then is a hack that ignores the genetic details of a particular trait and simply looks at the statistical patterns between parents and offspring. For experimenters, the crucial concerns are setting up breeding experiments that can explain how traits are inherited. For example, maternal half-sibs (offspring all born from the same mother) can eliminate the effects of the mother on the offspring. If the offspring all come from the same mother, then the mother’s genes will not explain the variation in the offspring. Only the population of fathers can supply the variation seen in the offspring.

Another approach to quantitative genetics comes straight from animal husbandry: if you want to breed animals that produce more of a certain product (the typical example is milk), then you can use equations to calculate how many cows to breed, and what their milk yields need to be, in order to produce a certain milk yield in the next generation. The difference in milk yield between offspring and parents is the “response to selection,” in this case artificial selection. This is the very idea of selective breeding that Darwin analogized to natural selection, so this same farming idea carries through to those of us studying the evolution of quantitative traits.

This last aspect is mainly my interest in quantitative genetics. We can use the equations I described in the last paragraph, and iterate them to simulate or describe the evolutionary process. The equilibrium solutions to these equations can tell us what traits are likely to evolve, all without going into the genetic details. Quantitative genetics involves a large number of approximations — skipping over the details — and that’s the nice thing about it mathematically. Despite using so many approximations, the equations are usually quite accurate.

The most recent development in quantitative genetics actually involves going beyond those approximations and mapping out actual genetic loci that are involved in the inheritance of quantitative traits. This is called Quantitative Trait Locus (QTL) analysis, and is also being called QTN analysis, for “quantitative trait nucleotide.” The meaning of the “N” should tell you just how specific some of these studies are getting. There is a controversy about this, however, since it appears that many studies from humans and animals show quite conclusively that the “hack” of quantitative genetics may not be so much a hack as a scientific reality. Genome-wide association studies and others that look at the whole genome are starting to show that indeed many loci of small effects are a better explanation than few genes of large effect in many cases, including human height and schizophrenia. This is especially weird, since Fisher’s “infinitesimal hypothesis” was only meant to solve a very specific problem, and probably was not meant as a real scientific hypothesis. What does this say about the operation of science?

I hope to meet some of you there next week. Thanks for reading.

Matthew V. Rockman (2012). THE QTN PROGRAM AND THE ALLELES THAT MATTER FOR EVOLUTION: ALL THAT’S GOLD DOES NOT GLITTER Evolution, 1-17 DOI: 10.1111/j.1558-5646.2011.01486.x

The Handicap Principle


, , , , , ,

Researchers use sexual selection theory to attempt to explain traits that are exaggerated, seemingly unrelated to survival and seemingly costly. I call these characters “ridiculous.” Almost every sexual selection talk starts with a collage of absolutely ridiculous-looking sexual ornaments and armaments. Most biologists have this “biological diversity slide” near the beginning, but other biologists have theirs filled with perfectly sensible looking animals. Animals that are clearly built for survival. Sexual selection, on the other hand, seeks to explain things like this:

Male Blue Peacock in Melbourne Zoo, Australia.

Male Blue Peacock in Melbourne Zoo, Australia. (Photo credit: Wikipedia)

One way to explain these extravagances is not to use sexual selection at all, but to say that it is, indeed necessary for survival. In fact, the whole idea of sexual selection arises out of a Bob McGuire argument: “Come on, that can’t possibly be necessary for survival! Chuck, you’re crazy.” Bob McGuire was a timeshare salesman who tried to talk Dr. Adamson and I into buying by saying things like “Come on, you’re not gonna take a baby camping, come on!” These appeals to disbelief are rather common in science, not just in anti-science.

The alternative looks at how ridiculous-looking traits do impact survival: if they are really costly, perhaps that cost relates to their value as signals. Perhaps they tell females something. Perhaps the cost itself is really important, and females should pay attention to that cost. How would a signal convey all that?

These important things to consider are fitness components. We talk a lot about fitness, but there’s no one measurable quantity that actually is fitness. Survival forms one fitness component, attractiveness forms another. When females want to have attractive offspring, one way to get them is just to mate with an attractive male: “You mate with a good-lookin’ bird like me, you’ll have good-looking babies, and good-looking grandbabies and so on. It’s a win!” The trait doesn’t need to be particularly costly to tell females that their offspring will be attractive. But what if the offspring don’t survive to mate?

Handicap to the rescue!

A really costly signal, on the other hand, could tell females that a male not only is attractive, but he’s able to survive well. The signal tells females that males carrying them are able to survive well because their trait does not impact their health as much as it would someone who was less healthy. This is called the handicap principle: if a male shows a handicap, he must be well-adapted and healthy, or else he wouldn’t be able to handle the cost. This idea was first proposed independently by Bob Trivers and Amotz Zahavi, and roundly rejected as completely preposterous. The 1989 edition of The Selfish Gene contains the typical argument: if it’s costly, then it’s costly and it will be selected against.

However, in 1990 Alan Grafen showed that handicap signals are evolutionarily stable: if everyone in the population is using a costly signal, then someone using a non-costly signal to convey the same information can’t make a living. If a signal is truly costly, then you can’t fake it. Grafen showed with his characteristic style that this applied to communication across the board, not just in sexual signals. Some researchers go so far as to say that every signal is a handicap, although I think “the finger” is enough of a counterexample. Incidentally, “the finger” did originally have meaning: it meant you had never been captured in war and were still able to shoot an arrow. I think this meaning has been lost since the Battle of Agincourt.

Kinds of handicap

There are a few different ways to have a handicap. One is called “pure epistasis” or “Zahavi’s Handicap.” In this case, every male grows the same trait, but it kills less-healthy males more often than it kills more healthy males. This is the version that Zahavi came up with originally, and it was widely ridiculed. It turns out the ridicule was partially correct, and this sort of handicap doesn’t really work. Zahavi’s handicap doesn’t work because after selection for survival, all males will start to look pretty much the same, and then females have no incentive to choose among males (a female who mates randomly will get the same good genes as one who goes to the trouble of choosing). Then the trait is useless, and costly, and will be lost from the population.

The second way is the “revealing handicap.” In this case all males grow the trait, but the signal itself has much better quality in more healthy males. The caricature is of a peacock’s tail: all males grow similarly-sized tails, but less healthy males let theirs drag on the ground and get infested with mites. It doesn’t kill them, but it’s not as shiny and dazzling and sexy as it would be for a healthy male. In this case the signal itself tells females how healthy a male is. This kind of handicap can lead to an exaggerated male trait, as long as there is sufficient genetic or environmental variation for the tail dragging on the ground.

The third way is condition-dependent signaling: males who are of good health grow larger traits and are less impacted by the costs of carrying the trait. The nice thing about this theory is that everything seems to work as far as genetic variation. Females benefit by mating with showy males, from having more healthy male and female offspring. Also, if the trait is condition-dependent, there is plenty of mutation in “condition,” which is basically the sum of all the selective forces across the genome. There should always be plenty of genetic variation in condition, so females should always benefit from being choosy.

I study how condition-dependent signaling plays out over the lifespan. If a male is really healthy, he can expect to live a long time, and he could potentially conserve his resources until he gets older. By growing his trait over a long period of time, he would have more opportunities to mate, and be just as sexy in middle- or old-age when selection is less intense. This does present a few problems in how the traits would actually change over time. If selection is intense enough, a male might be killed even for having a small trait, and once he gets older, his sperm will be harboring more deleterious mutations. I’ll go into more detail about these in future posts.

How do you do evolutionary theory?


, , , , , ,

Ronald Fisher as a steward at the First Intern...

Ronald Fisher in 1912 (Photo credit: Wikipedia)

My most frequently asked question has got to be: what organism do you work with? This is a funny question because I don’t work with any particular organism. The funny thing that happens is that after I tell people I don’t work with a particular organism and describe my work, they say “Oh, so in what kinds of animals?” This could be sending me several messages: one is that people have no idea that research into theory exists at all. I find this explanation unlikely since everybody knows about theoretical physics (which actually exists, it’s not just theoretical). There’s even a TV show in which everybody’s favorite character is a theoretical physicist. The message I get, instead, is that it’s quite strange to think about a “biologist” who doesn’t work in a lab, in a field, or a greenhouse. What would such a person do? How can you be a biologist and do all your work without touching real animals or plants? This is the point in the conversation when I tell people that I’m a mathematician getting a degree in biology. Basically what I tell people is “I do math” or “I solve math problems.”

The cool part is I get to decide what those problems are. I’d like to give everyone a better idea of what I actually do in my research. It’s one reason I started this blog and so I plan to explain things about my general way of working over a few posts.

Theoretical questions, mathematical answers

What scientists basically do is solve puzzles. These puzzles are usually posed as questions like “how does this plant grow?” or “When (in history) did this gene originate?” or something else highly empirical. When you want simply an accurate description of the world, when you want “How does this work?” you can go and get an organism and describe it. You can collect data, and you can formulate various hypotheses about the patterns in the data, then test those hypotheses by looking at other patterns in the data.

I do basically the same thing, except the questions that I handle are of a theoretical nature. That means that basically what I’m asking questions about are those hypotheses themselves, not the patterns in the data. I have a basic question about how sexual signals evolve. “How do they evolve?” means I want to know how a population of organisms, with some individuals carrying expensive organs or behaviors, would change over time. If a gene for such a signal arose, would most members of the population have it at some later time? Or would it go extinct? What does our existing theory tell us? How can I extend the theory we have so that we better understand this?

Sexual signals and behaviors appear to be costly, and so Darwin reasoned that they couldn’t come about by natural selection. He solved that puzzle by supposing a different form of selection that worked by mate competition, instead of simple competition to stay alive. Darwin was the first evolutionary theorist!

What Darwin didn’t have was a mathematical theory of how populations evolve. In evolutionary biology, we’re lucky to have such a theory, developed by the “Big Three” of Sewall Wright, Ronald Fisher and J.B.S. Haldane. Wright, Fisher and Haldane formulated the theory of population genetics as a mathematical description of how the frequencies of alleles change in populations over time. Just like Darwin, we don’t have to use math to answer these questions: we can take a guess or hypothesize, and go straight to testing that idea on real organisms. However, we can use mathematical tools to answer these questions, and mathematical answers are the most effective. There are several different ways to do this.

Some nuts and bolts

Everything I do can be called “mathematical modeling.” However, some of the time it is less mathematical, or more computational. If I am asking a question of how a gene frequency changes over time, then one way is the analytical approach: I write down a set of equations describing how the gene changes its frequency over time, and then I describe properties of those equations. The particular biological model I’m interested in will determine how the equations come out: all the information I need to say how the population will change should be in these dynamical equations.

The most important properties of dynamical equations are its equilibrium and the stability of that equilibrium. An equilibrium is a state (e.g. a gene frequency) that doesn’t change over time. So for example, in standard population genetics, if an allele’s frequency is zero, then it won’t change, unless there’s a mutation. Then zero is an equilibrium as long as I don’t allow mutation in the model. Stability means that when a population is close to equilibrium, it moves toward equilibrium. This is just like rolling a ball down a hill: when it gets to the bottom, it stops. That’s a stable equilibrium: if you throw the ball back up the hill, it rolls back down. This is important because we expect to find these equilibria in nature: we expect populations to get to a certain point and stay relatively close, and that’s what we should be able to see much of the time in nature. This is incredibly simplistic, but we have to start somewhere.

Another kind of dynamical model is a computational or simulation model. In this sort of modeling we input certain calculations as a computer program. Instead of writing out equations for how the variables change over time, we let the computer iterate a set of calculations to simulate how they would change. I could have a very good idea of the life cycle of a theoretical organism, and I program each piece (birth, growing up, mating and death) into a program, and then calculate how the gene or phenotype frequencies in the population(s) change over time. There’s two reasons to do things this way: the model of the life cycle might be extremely complicated, so that each piece might be a paper in itself before I solve my big question; or, we might explicitly mathematically know that there is no simple solution to the problem. In that case using a computer helps us come up with an approximation to the solution.

Both the analytical model I described and the computer model describe change over time: they are dynamical models. I like doing things this way. However, there’s another way, which is simply to mathematically describe how things are at a particular point in a population, or to describe the way an organism works. The solution to such a model is usually to say what the optimal phenotype of an organism is, given certain conditions. Evolutionary game theory and foraging theory are examples of this “phenotypic” approach. These approaches can be used to answer similar questions, but they do tell you fundamentally different things in most cases. A phenotypic model cannot tell you how something evolved, only the circumstances under which you could expect it to evolve. Phenotypic models are very popularly interpreted in behavioral ecology, where the researcher’s task is finding out how a particular strategy affects reproductive success of individuals. It’s not necessary to care, in that case, how something evolved: it’s just there and bears questioning.

The biological interpretation comes in when we consider the parameters of a particular model. Usually these special quantities determine if an equilibrium exists, and if it is stable or not, in a particular model. For example a particular model might have a mutation rate parameter, or a parameter describing the rates at which males and females meet, or a parameter describing the intensity of selection against a trait. Models of disease transmission have parameters describing how often a virus has a chance to move from one person to another. Parameters usually correspond to real biological quantities that could be measured, and that’s how hypotheses get formed and tested.

What is this all for?

The real point of doing all this is for somebody, a field or lab researcher, to be able to apply these ideas. If I can come up with a satisfactory solution to a theoretical puzzle, another researcher can take that theory to form a hypothesis and then test that hypothesis using fruit flies, or peacocks or some other real organism. My favorite kinds of applications look for patterns across large groups of organisms, many different species, and try to apply these ideas very broadly. Let’s say this species supports the conclusions of a model, and so does its closely-related species, since they correspond to different parameter values. Most of the time, however, theories get tested one organism at a time, and the research either supports the theory, or fails to support it in some way. This is the process of refinement, which is a lot of fun. For example, a theory could be supported in fruit flies, but not in cockroaches, but we find out that cockroaches are not a good species to use for testing the idea. This reveals a deficiency in the idea, since it doesn’t include cockroaches. Then I get to do some more work, and write another paper.

Sexual selection via sensory bias


, , , ,

We study sexual selection mostly to understand the evolution of costly, extravagant traits in males, for example flashy colors. Another reason is that sexual selection may play a role in the formation of new species (speciation). I mostly study sexual selection by female choice, meaning that certain females may prefer these flashy males; one way this could happen is that females might prefer particular males simply because they are conspicuous, or they look like food. This is called sensory bias. The simplest example is the guppy Poecilia reticulata where the females display a preference for orange objects because that’s what they eat. Males mimic this orange color and attract females. This is interesting because natural selection directly sets up the female preference: even if the preference has a cost, natural selection still favors its evolution, because it helps females survive. This could make new species form because different perceptual biases could be favored in different places.

Recent Servedio Lab graduate Alicia Frame asked whether sexual selection can then pick up these preferences and exaggerate them beyond what natural selection calls for. Sexual selection here would be indirect, meaning that the preference itself is no longer acted upon, but sexual selection on the attractive male trait could make the preference stronger. I’ll explain the mathematical details in another post, but basically whenever there’s direct selection on a trait (selection favors a trait), there is indirect selection on other traits that correlate with it. In this case the traits (the signal and the preference) occur in different individuals, but it’s simpler to think of two traits on the same individual organism. A simple example would be size of the brain and size of the body in humans. If there was selection to have a bigger brain in early hominids, then individuals with bigger bodies would also be favored by selection.

Alicia and our advisor Maria used a newer version of a biologically realistic model: there are “yellow fish” and “blue fish” that live in either yellow or blue environments. Males that contrast with their background environment are more conspicuous and thus more easily spotted by predators and preferred by females. Indirect selection is caused by correlations between traits. To have a correlation, you need lots variance in both correlated traits. What Alicia and Maria found was that when all females are choosy (the preference is there because of natural selection), there is not enough variation in the male trait for indirect selection to strengthen the female preference. This was true even when the male trait mutated back and forth between conspicuous and inconspicuous color morphs, creating more genetic variance. They did find exceptions, but I find this general result really interesting: natural selection alone is most important in determining the strength of the trait. If there is any strengthening, making the male trait even more ridiculous and conspicuous, then it will have to be the result of direct benefit to the female. Indirect selection is just too weak.

Alicia published this paper in the Open Access journal Ecology and Evolution, so anyone can read it free of charge. Here’s the abstract:

Evidence suggests that female preferences may sometimes arise through sensory bias, and that males may subsequently evolve traits that increase their conspicuousness to females. Here, we ask whether indirect selection, arising through genetic associations (linkage disequilibrium) during the sexual selection that sensory bias imposes, can itself influence the evolution of preference strength. Specifically, we use population genetic models to consider whether or not modifiers of preference strength can spread under different ecological conditions when female mate choice is driven by sensory bias. We focus on male traits that make a male more conspicuous in certain habitats—and thus both more visible to predators and more attractive to females—and examine modifiers of the strength of preference for conspicuous males. We first solve for the rate of spread of a modifier that strengthens preference within an environmentally uniform population; we illustrate that this spread will be extremely slow. Second, we used a series of simulations to consider the role of habitat structure and movement on the evolution of a modifier of preference strength, using male color polymorphisms as a case study. We find that in most cases, indirect selection does not allow the evolution of stronger or weaker preferences for sensory bias. Only in a “two-island” model, where there is restricted migration between different patches that favor different male phenotypes, did we find that preference strength could evolve. The role of indirect selection in the evolution of sensory bias is of particular interest because of ongoing speculation regarding the role of sensory bias in the evolution of reproductive isolation.

Alicia is now working on a postdoc in toxicology at the EPA in Research Triangle Park.

Alicia M. Frame, Maria R. Servedio (2012). The evolution of preference strength under sensory bias: a role for indirect selection? Ecology and Evolution DOI: 10.1002/ece3.273

New Journal Publishing Models


, , ,

Really interesting stuff is happening this week in the world of scientific publication. Scientists and editors are trying out new publication models that will change science. This is important because science publication is science. Scientists rely on citing printed works to give credibility to their arguments, their data, and basic knowledge. Everything that scientists put in a new article has to be attributed to an already-printed article, be a new result, or be generally accepted by the community. This last provision doesn’t stop most papers I see from citing Charles Darwin within the first paragraph. Even when we all know who he was and what he said, citing a printed work still lends the most weight to any point a scientist makes.

Why is this true? Because there’s an ironically Darwinian process behind publication called peer review. Once an article has been accepted for publication by a journal, it has been reviewed thoroughly and revised significantly. If it’s published, then several knowledgeable people agreed that it was important and well-done. Here’s a basic run-down of the process: I do some research and I want to tell people about it; I submit that paper to a scientific journal’s editor; the editor gives it to an associate editor who chooses (usually) three reviewers who prepare detailed comments on everything from spelling to the validity and the overall scientific significance of the results; the editor then looks over those comments and decides whether the paper is right for the journal, is well-done, etc. Then he can make one of several decisions (there are many variations, but these are the most common): reject the paper, give the authors a chance to revise and then start the process again, accept it given that the authors do what the reviewers suggest, or accept it outright. The last option is unheard of — except that it happened to me once; that was crazy — but it’s important for understanding new developments in publishing.

The last part of the process is that if the author gets the chance to publish the article, he has to pay to publish it. This is the weird part of scientific publishing that I think many people are unaware of, or might have thought they misheard. Yes, scientists pay to have their work published.

Ideally the process of peer review should live up to the ideals that I outlined above, but there are a lot of people who think it is outdated, unnecessary or evil. There are other problems in the process, such as authors paying for work, and then there is the cost of disseminating the research, i.e. actually getting it to its readers. As of this week, scientists are now toying with every part of this process along with the help of the internet.

The biggest development over the past ten years has been Open Access publishing, which takes care of the last part of the process, the dissemination of research. Open Access articles are reviewed just like other articles, but they are offered free of charge over the internet to anyone who wants to read them. This is unlike most journals that are only available to people with a subscription, for example students at a university where the library subscribes. The biggest argument in favor of Open Access has been that citizens are being made to pay for research twice: federal grants (that originate from tax dollars) pay to do the research and publish it, and then taxpayers have to pay to read the work once it’s published. There are now thousands of Open Access journals and you can read them as easily as you are reading this blog. This is a good thing.

Moving back through the process, a new sort of non-journal is experimenting with publication costs. PeerJ is not really a journal, but promises peer-reviewed publication on the internet for a one-time fee for the authors: you pay only $99 for one publication a year, instead of up to $1000 for one publication at a time. PeerJ also promises to publish articles on the basis of scientific merit rather than impact. Most journals, including my favorites, will only accept paper if the editor and the reviewers can agree that it’s significant to the scientific community. PLoS ONE changed that by accepting papers only based on whether the science was well-done, and PeerJ plans to do the same thing. As you can imagine PLoS ONE publishes a lot of papers: there are over 2,000 a month these days, which makes it really hard to find interesting papers to read. If I publish a paper in PLoS ONE, I will definitely blog about it.

Dealing with the problems of peer review is the goal of Peerage of Science, which is a network of scientists that distribute their work and get it reviewed before submission to a journal. I think the idea here is that you can then submit your paper saying that it’s already been reviewed. I like the idea of forming a community of peers where we can review each others’ work without things becoming competitive. This is one of the biggest complaints about peer review. I’ve never experienced competition during the review process, but I know people who have thought that was going on. Reviews are typically anonymous, but Peerage of Science encourages the breaking of anonymity.

A new microbiology journal called mBio plans on dealing with the editor’s decision-making process. Instead of all the variants, and the reviewers getting whatever they want (“we kindly thank the reviewers for making us do additional experiments that had nothing to do with our hypothesis, but that seem to fit the research program of another biologist we know very well”), mBio promises to either reject a paper or accept it with minor revisions. A minor revision is something like the decision to not include a figure, or to add one or two additional citations. This means that the editor can make a faster decision, but it also means that reviewers are given less (more?) power. Sounds like if a paper needs significant revisions, mBio will just reject it, saving the authors some time.

I think what all these experiments are going for is making scientific publication a lot more like a wiki: a place where people can easily access each others’ work and data, easily share their work and data, and review is still there. Review, with all its problems, is still incredibly important. Lots of people have suggested alternatives, and there are some good ones, but I still appreciate the review process, even if I don’t like it all the time while I’m in the middle of it. I recently had a paper rejected because the reviewers just didn’t get what I was saying, but that is my problem, not theirs, so they did their jobs. Moving scientific publication in the direction of openness, through any of these ideas, is a huge step forward.

We should be a long way from the days when scientists would deliberately obfuscate their results from their peers — Galileo disseminated his results to other astronomers written in a cipher. The biggest thing holding us back is the idea that scientists own their research: I don’t believe I own my ideas, but there are people who want to keep their ideas to themselves as long as possible (i.e. until after it’s been through peer review). And of course there are people who profit from keeping scientific results hard to get at. However, as the free software movement shows, today’s technology (incidentally, built out of free software) makes sharing incredibly easy for those of us who want to share. The people who created arXiv understood this even when it wasn’t so easy.

Thanks for reading.

Sexual selection in humans: some interesting recent work


, , , , ,

Researchers have found sexual selection important in the evolutionary history of humans, and a lot of researchers are focusing on the roles of mate choice and life history in major transitions in human evolution. I find the transition from hunter-gatherer to agricultural civilization the most interesting. This week I’ve read three interesting papers on three interesting facets of human sexual selection. These studies also did things in three different ways: there is a study of psychological study of modern (i.e. living) human mating preferences, a study of pre-industrial humans using historical data and a theoretical study using mathematical and computer models.

The first study is a twin study of human mating preferences in men and women. Brendan Zietsch and Karin Verweij at the university of Queensland, and Andrea Burri from King’s College, London gave surveys to men and women in both monozygotic (MZ, “identical”) and dizygotic (DZ, “fraternal”) twin pairs, asking them to rank the qualities of mates that they found desirable. The traits were qualities like “kind and understanding,” “healthy,” “intelligent,” “good earning capacity,” “good housekeeper,” “wants children,” and of course “physically attractive.” The most amusing characteristic of their raw data was, of course ,the differences between the sexes. Members of both sexes valued “kind and understanding” the most, and in line with other studies, men valued physical attractiveness more than women did. The least important qualities to men and women were “religious” and “university graduate.”

The evolutionary part of this study was using the twins to find the ability of natural selection to affect these preferences. MZ twins have the same genetic material, therefore differences between them are due to the environment. We can get more information on the role of environment by looking at DZ twins, who share some genes. From these data the researchers quantified the amount of additive genetic variance, dominance genetic variance and environmental variance. When individuals differ in their survival or reproductive success and it’s because of additive genetic variation — the number of certain copies of particular genes — then all those differences in fitness are passed on to their offspring. Additive variation is therefore the raw material for selection. These researchers found that the mating preferences with the highest heritability (biggest differences in preferences due to genes) were those that ranked highest on people’s priority lists. The researchers also questioned why the heritabilities should be so low (around 20% in most cases): though they entertained many possibilities, I think the most promising is that selection is acting very strongly on women’s and men’s mating preferences right now.

That brings us to the second study, from a multinational group of researchers that was published in Proceedings of the National Academy of Sciences (PNAS). The data on marriage (i.e. mating), births and deaths was mostly compiled from church records, family bibles and tax records in pre-industrial Finland. Another really interesting church-record data set producing results on human evolution, was derived from records of the pioneering Mormons of Utah in the late 1800s. The researchers studying the Finnish population asked if they could determine whether natural and sexual selection was acting on this population over the time period of their data, and what the crucial periods in the lifetime were: what fitness components contributed to overall fitness? Was it survival to adulthood? Was it how many times you married? Was it how many children you had? They found that in this society, where serial monogamy (i.e. remarriage) was common, and extramarital affairs and divorce were severely punished, that survival to adulthood made the biggest difference. However, they also found that the raw material for sexual selection in this population: interestingly from my perspective, most men remarried younger women who could still produce more children.

They also found that sexual selection might be able to act on the ability to remarry in women as well as men in this population. Also interesting: the wealth of individuals (whether or not they owned land) was totally unimportant to any fitness component, either under natural or sexual selection. The authors of the paper emphasized that natural and sexual selection can still act in our species, despite the demographic changes that came with the agricultural revolution. This is an important finding, and their data is totally awesome.

The third study I read this week used mathematical models to study the possible transition from a promiscuous mating system in human ancestors to our more-familiar system with long-term pair-bonds. Sergey Gavrilets of UT Knoxville authored this study, also published in PNAS. This is a contentious area, with many sources of interest, including anthropologists, social theorists and evolutionary biologists. I’m glad to see that a theory paper is getting some popular attention, and particularly that anthropologists are paying attention to it.

Gavrilets studied four different models of how males can obtain mates, and how females derive their fecundity, at least partially from male behavior. He used these models to ask if there wa sa relationship between the fighting ability of males and how much they provisioned their mates: we often assume, as is standard in economics as well, that each organism has a finite amount of resources to devote to various activities, so he divided male activities into fighting versus something else. All these models led to a state where males did nothing but fight, and females had lower fitness than if they got some direct, material benefits (food) from their mates. This is a low-fitness state: good for males who can fight to gain more mates and thereby more offspring, but not so good for females, who could have more offspring and survive better if the dudes would just cut it out.

Then Gavrilets added two wrinkles: a negative correlation between male fighting ability and mate provisioning, and female variation in faithfulness to their mates. Using a computer model, he then simulated evolution to show that populations would move more toward monogamy and long-term pair bonds. Gavrilets’ conclusion is that low-ranking males (not as good at fighting) would could increase their reproductive success by provisioning their mates; females can reward and reinforce this by not fooling around, which would force these males to provision offspring who got genes from an aggressive, promiscuous male.

All these studies show some really interesting ways that sexual selection works in humans, and may have played a role in our past. There is male and female mate choice going on. There are factors other than simple mate choice, for example in Gavrilets’ models. There are also life-history factors: as the Finnish study showed, survival to adulthood is a necessary prerequisite for mating success. Also, knowing that sexual selection can still act in humans, who are mostly monogamous, is really exciting.

A few caveats are in order, however: all of these studies emphasize selection. So do I, since I study selection. However, it’s easy to get caught up in the idea that selection is the evolutionary mechanism, when in fact there are other forces that are potentially much more important, particularly genetic drift, which randomly causses alleles to disappear from populations. We often think of genetic drift as something that produces variation between populations: the differences in appearance between Europeans and Asians, for instance, are conceivably due to drift. But who wants to think that critical parts of our identity — for example our mating behavior — could be due to a process even more stupid than natural selection? Drift is not just stupid, it’s stochastic — even worse! In other words, selection at least has the appeal that “it’s not totally random!” as biologists are often heard to say to religious zealots. Drift, on the other hand, is completely random. Not very pretty. The trap of thinking that everything interesting is due to selection is called adaptationism.

The other caveat is that most studies of humans assume that our current monogamous mating system is derived, in other words a recent adaptation. Most studies I hear of, be they from anthropologists, psychologists, or evolutionary biologists, assume that our ancestors were promiscuous or polygynous. This is intuitively appealing for scientific reasons — men are, on average, larger than women — and for social reasons — we like to think of ourselves as new, developed, derived and interesting. Whatever we are doing right now is often seen as a good thing, and we know that in the past what those people did was not a good thing. However, I have yet to see any data that supports this idea. The specific significance of Gavrilets’ paper hinges on the idea that our ancestors were not monogamous. However, this could be a good case of The Platypus Fallacy: just because gorillas and chimpanzees have different mating systems from modern humans does not mean that our ancestors did.

Thanks for reading.

  • Zietsch BP, Verweij KJ, & Burri AV (2012). Heritability of preferences for multiple cues of mate quality in humans. Evolution 66 (6), 1762-72 PMID: 22671545
  • Alexandre Courtiol,, Jenni E. Pettayd,, Markus Jokelae,, Anna Rotkirchf, and, & Virpi Lummaaa,b (2012). Natural and sexual selection in a monogamous historical human population PNAS DOI: 10.1073/pnas.1118174109
  • Sergey Gavrilets (2012). Human origins and the transition from promiscuity to pair-bonding PNAS DOI: 10.1073/pnas.1200717109

The social environment as a source of variation


, , , , ,

Some of the most interesting studies on life-histories and mate choice have been done with laboratory populations of insects. Crickets are especially well-suited to these studies because their most salient characteristic, their call, has qualities that are affected by different evolutionary forces. Calls have a frequency (pitch), but males also exert effort calling. Several researchers have picked up on the different meaning of these two variables for sexual selection in crickets. The “ears” of female crickets are probably tuned to a particular frequency, just as (American) human ears seem tuned to 440Hz. This should lead to stabilizing selection: males whose calls are too high-pitched will not attract mates or attract the wrong ones (e.g. from another species); the same should apply to males whose calls are too low-pitched. However, frequency should be mostly independent of the amount of time a male spends calling: this should depend on his health (how much time can he spend?), the predators and parasites that might tune in to his call, and the density of other males, i.e. competitors. The overall effect should be directional selection, i.e. favoring males who call more when they can (condition-dependence).

Michael Kasumovic and his colleagues Matt Hall and Robert Brooks recently raised male Australian Black Field Crickets to determine if their juvenile environment affected their calling effort and their pitch. The researchers ingeniously hypothesized that perceived density of competitors would affect male song in different ways: perceiving more competitors should not affect pitch, whereas competitor density should definitely affect calling effort. If males perceive that they have more competitors, they will have to call more to find mates. This produces directional selection on calling effort. However, females will still tune in to the same frequency, regardless of how many males are around. The effect of juvenile social environment therefore coincided with whether directional selection or stabilizing selection was the dominant force on a particular trait. They fooled the crickets into believing there were more competitors by broadcasting calls to developing males in the lab. Since they also knew that female choice was affected by the females’ perception of male density, they also raised females under similar conditions.

Crickets mating

I find this study really interesting because it shows how many possible sources of phenotypic variation there are. When we consider the lek paradox, a big problem is that the conditions under which we expect to lose genetic variation are very narrow: we have to suppose that the genes act in a certain way, that there is no mutation, and so on. We might as well deal with friction-less pulleys and billiard balls. My research focuses on phenotypic variation due to age, and Kasumovic and his colleagues have focused on the social environment. This study reminds us that the complexity of life means that there is a huge number of reasons we should expect lots of phenotypic variation. That phenotypic variation should lead to plenty of ways that we can maintain genetic variation in populations. However, science proceeds in baby steps of understanding: each potential idea has to be tried out and tested to death. We can think of these sources of variation a lot faster than we can do experiments, or even produce worthwhile theory (that is, find out if the ideas really make sense). The lek paradox will therefore be with us for a while.

Another really interesting thing about this article is that it’s published in the new Open Access journal Ecology and Evolution. That means everybody interested can go and read it: while you’re there check out all the other primary research articles you have access to free of charge. This is primary research, i.e. the research papers written by the people who performed the experiments.

Here’s the abstract:

The juvenile environment provides numerous cues of the intensity of competition and the availability of mates in the near environment. As research demonstrates that the developing individuals can use these cues to alter their developmental trajectories, and therefore, adult phenotypes, we examined whether social cues available during development can affect the expression and the preference of sexually selected traits. To examine this, we used the Australian black field cricket (Teleogryllus commodus), a species where condition at maturity is known to affect both male calling effort and female choice. We mimicked different social environments by rearing juveniles in two different densities crossed with three different calling environments. We demonstrate that the social environment affected female response speed but not preference, and male age-specific calling effort (especially the rate of senescence in calling effort) but not the structural/temporal parameters of calls. These results demonstrate that the social environment can introduce variation in sexually selected traits by modifying the behavioral components of male production and female choice, suggesting that the social environment may be an overlooked source of phenotypic variation. We discuss the plasticity of trait expression and preference in reference to estimations of male quality and the concept of condition dependence.

Remember this article is available to everyone for free, so please go read it and learn more about the authors and their other interests.