Microevolution Study Guide PDF

Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. Creating variation: When a new mutation occurs within a gene/locus, we can call this a new allele. The genotype for a diploid organism relates to both alleles at each gene/locus. Once a new alle...

Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. Creating variation: When a new mutation occurs within a gene/locus, we can call this a new allele. The genotype for a diploid organism relates to both alleles at each gene/locus. Once a new allele has arisen via mutation, it can also be introduced into other populations by migration. Migration results in gene flow if/when the individual(s) arriving in the new population are able to establish and produce offspring. New genotypes (but not new alleles) can be introduced into a population via recombination. The majority of mutations are either neutral or deleterious. Nonsynonymous mutations (i.e. those that lead to changes in the encoded amino acids) are particularly likely to be deleterious but can occasionally confer a fitness advantage. Synonymous mutations (those that do not alter the encoded amino acids thanks to redundancy in the genetic code) can still impact fitness via their effects on gene expression and protein structure. Gene flow increases genetic variation within populations, and a group of subpopulations that are linked by migration are referred to as a metapopulation. As a result of migration, these metapopulations are more genetically similar than you would expect by chance for populations that are isolated from one another. Recombination creates offspring that can have different allelic combinations than either parent, leading to more phenotypic variation upon which selection can act. Importantly, however, alleles are broken up by recombination, making selection for a particular genotype (e.g. Aa) slower, as offspring of the less fit genotypes will still be produced. Overall, however, increased standing genetic variation (i.e. within-population variation) allows an evolutionary response to selection to be more efficient. Genotype to phenotype: An organism’s genotype is the set of genes that it carries. An organism’s phenotype is all of its observable characteristics — which are influenced both by its genotype and by the environment. There are many reasons that we are yet unable to accurately predict what an organism will look and act like based on its genome alone, including because of genetic interactions among loci (epistasis), DNA methylation and chromatin accessibility leading to altered gene expression, or as a result of alternative splicing. In this latter case in particular (where exons from the same gene are joined in different combinations, leading to different, but related, mRNA transcripts), the same exact gene/allele can result in different proteins with distinct structures and functions! Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 1 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. Because gene expression and regulation (i.e. if and how much a gene results in mRNA and then protein) can be so dependent on the environment that an organism is experiencing, there are very often environmental effects on the phenotype that can not be (easily) predicted from the genotype. For some traits, the environment shapes the trait independently of the genetics of the organism. For other traits, the genetics are the primary driver of the trait and the environment has little impact. And for other traits yet, there is a genotype by environment interaction, in which different genotypes respond to the environment in different ways. In cases where the phenotype is influenced by the environment, we expect less rapid evolution in response to selection. In the most extreme case of a trait that is entirely shaped by the environment, there can still be strong selection on that trait (e.g. only large individuals survive) but we would not expect a response to selection. Population Genetics: Population genetics is the study of the genetic composition of populations, including distributions and changes in genotype and phenotype frequency in response to the processes of natural selection, genetic drift, mutation and gene flow. The idea is that one can conceptualize the transmission of alleles across generations, the selection acting on them, and the importance of population demography (e.g. size, bottlenecks, etc…) to both understand and predict how allele frequencies will change over time. The cornerstone of many of these models is the idea that randomly mating populations of sexually reproducing diploid organisms will have alleles (mutational variants at a given locus/gene) that are in Hardy-Weinberg equilibrium. The H-W equilibrium suggests that, in the absence of mutations, natural selection, nonrandom mating, genetic drift, and gene flow, genetic variation in a population will remain constant from one generation to the next. Moreover, at this equilibrium, you can determine the genotype frequencies in a population based on the frequencies of each allele. For simplicity, we focus on a locus with two alleles (A and a) but the equilibrium principle holds with as many alleles as you can keep track of. The basic idea is that each parent contributes one allele at each locus to their offspring (because gametes are haploid) and thus would can calculate the probability that an individual in the next generation has a given allele based on it frequency in the parental population. In other words, if a is rare in a population, the chance of an aa offspring genotype is low; the chance of an Aa genotype is not quite as low, and the chance of an AA offspring genotype is high. You can visualize this as a Punnett square (see below), where the probability of ‘grabbing’ to A alleles from the allele pool by chance is p2, the chance of ending up with two a alleles by chance is q2 and the chance of receiving one of each (and becoming a heterozygote is 2pq. The sum of the entries in the below Punnett square is p2 + 2pq + q2 = 1 And the allele frequencies also always sum to 1: p + q = 1 Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 2 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. Similarly, you can calculate the frequency of each allele at any given time as: ft (A) = ft (AA) + ½ ft (Aa) = p ft (a) = ft (aa) + ½ ft (Aa) = q And remember that to turn number of individuals of each genotype into frequencies of that genotype, you have to do a bit of math Genotype AA Aa aa Total # individuals 48 384 768 1200 Frequency =48/1200=0.04 =384/1200=0.32 =768/1200=0.64 1.0 Based on this, you can ask if the population is in H-W equilibrium, for example, by calculating the square root of ft (AA), which is 0.04, to get p. In this case, p=0.2. Because under H-W, p+q=1, we can infer that q=0.8. We can then test if the population is indeed in H-W equilibrium by calculating the expected ft (Aa) which is 2pq = 2 x 0.2 x 0.8 = 0.32. This matches the ft (Aa) that we calculated from the actual number of genotypes in the population, so we can conclude that the population is indeed in H-W equilibrium. When we say this equilibrium could be perturbed by selection, we mean that if one genotype has higher fitness than another – we expect more of the alleles from that genotype to be represented in the next generation, moving the population away from this equilibrium. We represent fitness as W and can assign each genotype its own fitness and then ask how that fitness compares to the average fitness of the population, which is: w= p2wAA+2pqwAa+q2waa Once we have an idea of the relative fitness of each genotype we can then predict the allele frequencies in the next generation (p’ and q’): 1 𝑝2 𝑤𝐴𝐴 + 2 (2𝑝𝑞𝑤𝐴𝑎 ) 𝑝′ = 𝑤 and ′ 𝑝2 𝑤𝐴𝐴 + 𝑝𝑞𝑤𝐴𝑎 𝑝 = 𝑤 With this, we can calculate the change in allele frequency between two time points as: ∆𝑝 = 𝑝′ − 𝑝 Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 3 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. 𝑝2 𝑤𝐴𝐴 + 𝑝𝑞𝑤𝐴𝑎 𝛥𝑝 = −𝑝 𝑤 Based on this we can conclude: - If Δp > 0, then natural selection has led the A allele increasing in frequency - if Δp < 0 then selection has led the a allele increasing in frequency - If Δp = 0 then no allele frequency change has occurred (equilibrium). Remember that fitness is relative, because selection will only select for a beneficial genotype if that allele exists in the population. In other words, if aa is the most fit genotype that could exist, but there are no a alleles in a given population, selection cannot favor this allele and evolution will not happen. This will remain true until the allele is introduced into the population by mutation or migration. Mutation-selection balance One reason that variation exists within a population at any given time is that mutation is constantly introducing variation, even if most of this new variation is detrimental or neutral. Mutation–selection balance is an equilibrium point at which the rate in which deleterious alleles are created by mutation equals the rate at which deleterious alleles are eliminated by selection. Just like each genotype has a fitness, we can think about each genotype having a selection coefficient that acts on it (s). We calculate s by subtracting each fitness value from 1.0 s = 1- w Remember that a larger s means stronger selection against the allele. Because we consider the maximum fitness to be 1 (i.e. producing as many offspring as possible), we can therefore say that the maximum value of s is also 1. A genotype with maximal fitness will have a selection coefficient of 0 and should not be selected against. A genotype with a fitness of 0 will have an s of 1 and will be strongly selected against (we can think f this as a lethal mutation). So we can think about this balance as a constant tug of war between selection removing deleterious alleles and mutation (re)introducing them captured by q2 = μs where μ is the mutation rate at that locus. In the case of populations that are connected by migration, there will be a similar push and pull between selection and migration – especially if the allele is beneficial in one population but detrimental in the other. If we want to predict how selection will act on a given genotype (remember that selection doesn’t act on the allele) we can use the following formula: Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 4 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. 𝑞2 (1 − 𝑠) 𝑓(𝑎𝑎) = 1 − 𝑠𝑞2 In this case, if s was 1 we expect all aa offspring to be killed or otherwise blocked from contributing alleles to the next generation. However, we do not expect the a allele to be lost from the population. This is because as selection reducing q in the population, the number of aa genotypes will decrease but the allele will remain in the heterozygote state. In the case that a is recessive (i.e. that the 𝑤𝐴𝐴 = 𝑤𝐴𝑎 ) then the allele remains hidden from selection in the heterozygote state. For a deleterious recessive mutation, we expect that the strength of selection (the selection coefficient, s) will increase the speed at which it is lost, but unlike the formula above suggests, if you want to predict q’ as opposed to 𝑓(𝑎𝑎), you need to take into account the frequency of Aa genotypes in the population (2pq) as you see in the equation to the left. You will also notice in the figure to the left that under all s values, there is a very long tail showing just how inefficient selection is at removing a deleterious recessive allele entirely. Balancing selection In the case where the heterozygote genotype has even higher fitness than either parent, we expect both alleles to be maintained in the population over time. This is true even if one (or both) of the homozygous genotypes are detrimental, as is the case in Sickle cell disease. In this way, a heterozygote advantage (shown below) can maintain a polymorphism within the population. Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 5 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. In a similar way, variation can be maintained over time as a result of negative frequency dependent selection. In this case, the fitness of a genotype is higher when it’s rare in the population. But as selection for this genotype results in evolutionary change, the genotype will become more and more common and its fitness will decrease. As such, you can maintain many alleles because each will be selected against as soon as it has been successfully selected for. Note that, just like other forms of selection, these forms will lead to violations of the H-W equilibrium. Artificial selection and breeding In the case the alleles/genes underpinning a trait are not known (which is the case for the majority of traits we study), there is still a useful way to predict if/how evolution will proceed. The Breeder’s equation is a useful framework for thinking about evolution – particularly in the case of artificial selection (as opposed to natural selection) and/or when the trait is likely polygenic (i.e. has many genes/loci shaping it). R = h2 S In this equation R is the response to selection, h2 is the narrow-sense heritability of the trait and S is the selection differential (not to be confused with the previously discusses selection coefficient, s). Remember that for this equation to be useful you need to know 2 or the three variables, and then you can calculate the third. R can be usefully calculated as the difference between the mean of the population before selection and the mean of the offspring population. Note that if the trait is entirely heritable (h2=1), you would expect a response to selection that is exactly equivalent to the selection differential (S) such that if you selected only 7 foot wheat varieties to set seed in the next field, you would end up only 7 foot wheat plants. S can be usefully calculated as the difference between the mean of the population before selection and the mean of the population that reproduces (in the case of artificial selection, the mean of individuals you select to reproduce). Trait heritability can be estimated in two ways: First, if you run your breeding trial and know R and S, you can figure out h2 based on this: h2 = R / S. So if the population made it halfway between the starting phenotypic mean and the selected mean, you would conclude that h2 = 0.5. The other way you could calculate heritability is using a parent-offspring regression. Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 6 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. In this case, the slope (m) in the equation for a line y=mx+b gives you your estimate of the heritability. In other words, if there is a perfect correlation between the average trait value of the parents and the average trait value of the offspring, the slope would be one and you would conclude that the trait is entirely heritable. Genetic Drift Genetic drift is when allele frequencies change across generations is driven not by selection but by randomness/chance. If you think back to H-W equilibrium, a new genotype (offspring) is created by random sampling of alleles from the parental population (assuming random mating) and this is predicted by the frequency of the alleles in the population because it is all driven by probabilities. As a population moves from “infinitely large” to smaller, stochasticity comes into play such that you could randomly get two rare alleles together in a genotype more often than you’d predict. In this case, an allele that is not under selection could change in frequency and eventually either be lost of fixed in the population. In the case of a newly arisen mutation this means that is far more likely to be lost by drift (i.e. by chance), but occasionally it will become more and more common. Mathematically, we can predict that in population of N diploid individuals, the chance of fixation of a new mutation is 1/(2N) and the chance of loss is 1 – (1/2N). This demonstrates why genetic drift is most important as a force in smaller populations, given that it relies on stochasticity/random change, and these are more pronounced in smaller populations. It is also common as a result of population bottlenecks, where there is a period of time where far fewer individuals reproduce and thus contribute to the next generation. During such bottlenecks, we predict genetic diversity to be lost but also the diversity that remains to be a random (and potentially randomly skewed) proportion of the original diversity. Finally, genetic drift is an important force in part because it separates the genetic diversity of populations over time – especially if/when they are not connected by regular gene flow. Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 7 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. While drift decreases genetic diversity within populations over time (as a result of random loss of alleles), it increased diversity among populations. Population demography and diversity As we have discussed across numerous lectures, the size (and fluctuations in size) of a population as well as its connectedness to other populations via migration both shape the evolution of that population in meaningful ways. One useful way to think about the impact of population size is to characterize the effective population size (Ne), which is the actual number of individuals within one generation contributing alleles to the next generation, as opposed to the census population size, which is the number of individuals present within a population (reproducing and non-reproducing). Quantitative genetics Most traits do not neatly fit into the single gene, two allele model, but that does not mean the population genetic approach is useless. For many traits with multiple genes underlying them (so-called polygenic traits), they can still behave – for all intents and purposes – like a single locus, two allele model. In other cases, however, it is important to consider the multi-gene effects in order to predict the evolution of a trait. Quantitative traits are those that vary continuously within the population, usually approximating a normal distribution among individuals (with a mean and variance). However, you can have polygenic traits that are meristic (can be counted) or dichotomous (threshold) as well. Overall, we can loosely predict that the more genes are involved in explaining a trait, the more variation there is likely to be in that trait (although remember that environmental and GxE effects can also generate what looks like genetic variation but it not heritable). The key advancement of quantitative genetics was the idea that to map genotype to phenotype in a useful way, you need to partition the variance in the phenotype (VP) into that which is explained by the genetics (VG), that which is explained by the environment (VE), and that which is the result of GxE interactions (VGxE). VP = VG+VE+VGxE Using VG, we can calculate the broad-sense heritability of a trait H2 (as opposed to narrow sense heritability, h2, which we have already discussed). H2 is the variation in genetic variation / variation in phenotype or H2= VG / VP. This heritability is important to consider for clonal, asexual species (no recombination), where genes that are found in a genome are likely to remain together in the offspring such that they may as well be a single locus. Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 8 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. The second advancement is the idea that even the genetic variance needs to be partitioned into that which is inherited but not heritable and that which is truly heritable. In this case, we can think about the role of genetics in shaping a trait as that which is due to additive genetic variation (VA), Interaction (epistatic) genetic variation (VI), and Dominance genetic variation (VD). VG = VA +VI +VD VA is the additive genetic variation underlying the trait and is the substrate for natural selection on phenotypes. This is because each allele directly contributes to variation in the phenotype, regardless of the effect of another allele on the phenotype (epistasis or dominance). In other words, the effects appear in the next generation even though recombination and independent assortment are occurring. VI is the genetic variation that is caused by inter-allelic gene interactions/Intergenic interaction. These are the genetic interaction between two or more genes located in the same or different chromosome(s). Two or more gene products are involved in the outcome of one common trait. From the perspective of evolution, because these diploid combinations of alleles are broken up each generation, they are not transmitted in haploid gametes, so this variation is not heritable in a narrow sense. VD relates to genetic interactions within a locus/gene (i.e. it represents the Intra-allelic gene interactions). These are interaction between the two alleles of a single gene (Aa genotype might have a different phenotype than AA genotype). In other words, the effect of an allele on the phenotype at one diploid locus depends on the identity of the other allele (from the other parent). Under codominance, AA, Aa, and aa each give a different phenotype (e.g. red, pink, and white flowers). From the perspective of evolution, because these diploid combinations of alleles are broken up each generation, they are not transmitted in haploid gametes, so this variation is not heritable in a narrow sense. Remember that to predict how a trait will respond to selection, we have to partition the variation of that trait into the variation explained by the environment and the variation explained by the genetics. We can use VA and VP to estimate narrow-sense heritability (h2). Narrow-sense heritability = Variation in additive genetic variation / Variation in phenotype or h2= VA/ VP. Importantly, if a trait is polygenic but every gene is acting independently and simply acting in an additive fashion to shape the final trait value, this would still be a highly heritable trait in the narrow sense and the VG would be additive! On average, the more genes are involved in the expression of a particular phenotype, the narrower phenotypic categories become, and the more continuously distributed a trait is. But how can you know how many genes there are? There are (at least) two useful approaches. The first is the QTL approach, where you genotype and phenotype large numbers of individuals derived from experimental crosses between two individuals from the opposite extremes of the phenotypic spectrum for your trait of interest. You can use the Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 9 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. variation you create through continually crossing the offspring (thus breaking up linkage among genes that are physically linked on a chromosome thank to crossing over/recombination during meiosis) to look for correlations between alleles at particular loci and your phenotype. Based on how many linkage groups (groups of genes that remain linked because they are physically close together and have not yet been broken up by recombination) you can begin to predict whether there are multiple loci underpinning your trait. Fitness Landscapes Correlated responses to selection are used in quantitative genetic studies to describe changes that occur in a trait as a consequence of selection imposed on some other trait or traits. It arises because the traits are genetically correlated (for example they are encoded by the same gene, or the genes underlying one trait interact with the genes underlying another). This is important because it can lead to critical trade-offs, such that evolution in response to one selection pressure affects the response to another selection pressure. For example, imagine a case where the evolution of resistance to one pathogen alters the immune system in a way that increases sensitivity to another. These correlations matter in evolutionary biology because they can constrain how organisms respond to one selection pressure in light of others. A fitness landscape (or adaptive landscape; left) is a way of visualizing and/or imagining how populations might evolve in response to selection based on the relative fitness of all possible genotypes. To the left you seen an example visualization of a two dimensional landscape (the most common type you will see), where the arrows represent various mutational pathways that the population could follow while evolving on the fitness landscape. Here, the fitness of each genotype is the height of the landscape (i.e. shown on the Y axis). Genotypes which are similar are said to be close to each other, while those that are very different are far from each other. The set of all possible genotypes, their degree of similarity, and their related fitness values is then called a fitness landscape. The key point to remember about the fitness landscape idea is that it demonstrates how/why a population might not be at its ultimate fitness (e.g. that are some possible combinations of allele that are currently inaccessible or unlikely to evolve). Imagine two scenarios: (i) the fitness landscape looks like a single volcanic island. In this case there are mutations that increase fitness and those that decrease fitness. Each will be, Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 10 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. in turn, selected for or against based on these effects and we predict the population to make a slow march up the mountain until it reaches the top (at which point all mutations are predicted to be deleterious. (ii) The fitness landscape is made of many peaks and valleys (as shown above). In this case, a population that is at one of the lower peaks, by chance as a result of the genetic variation that arose via mutation or migration earlier in their evolution), will be unlikely to evolve towards a higher peak because to get there it would have to cross a valley of low fitness. In other words, you would first need to have relaxed selection against mutations that are deleterious (moving the population into a valley) in order to be able to start climbing the other peak. Study question: based on what you learned about epistasis among alleles, can you explain how epistasis might create a fitness ‘valley’? Importantly, there are indeed ways that populations are predicted to cross from one fitness peak to another. Any external factor that reduces the strength or efficacy of selection can move populations off of their fitness peak, potentially allowing them to start climbing a different one. For example, population bottlenecks can suddenly move the mean fitness of a population far from its peak based on the random nature of who survives (or migrates) and/or how much genetic variation there is left after the bottleneck event. The other key point to remember about fitness landscapes is that they are not (always) static! Because fitness is relative (to the environment but also to the other genotypes in the population), these landscapes can shift as a population is evolving. The Evolution of Sexual Reproduction The two-fold cost of sexual reproduction takes into account two ‘problems’ with sexual reproduction as a way of creating offspring: (i) in a sexual population, the non-offspring bearing sex only contribute genes, and not (always) resources, to offspring, making them essentially a "cost" to the offspring bearing sex’s reproductive output. All else being equal, an asexually reproducing individual (capable of directly giving birth to offspring) will produce twice as many offspring-bearing offspring and four times as many grandchildren as a sexually reproducing offspring-bearing (female) individual. This has been dubbed “the cost of males.” (ii) in comparison to asexual reproduction, where an organism can produce offspring that are genetically identical to itself, sexual reproduction results in only, on average, 50% of an individuals genes being passes on to offspring. Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 11 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. Together, these costs lead to an evolutionary paradox! Sexual reproduction should not exist and yet sexual reproduction is observed to be prevalent in most species, suggesting that the benefits of genetic diversity and adaptation outweigh these two-fold costs. One solution to this problem might be that sexual recombination increases genetic variation within populations, thus increasing the evolutionary response to selection. Remember from quantitative genetics that linkage between loci is broken up over time thanks to both independent assortment and crossing over events during meiosis. This means that sexual reproduction/ recombination can break up two mutations/alleles that would otherwise be stuck together in the same genome. This process speeds up the efficiency of natural selection because it allows one allele that may be beneficial to be separated from a deleterious mutation that it happened to occur nearby to. Thought experiment: Imagine a new mutation that gives its host organism a mildly beneficial fitness advantage in the local environment. What happens if this mutation happens to occur in an individual that is otherwise maladapted to its environment? How could sexual recombination change the speed at which this allele spreads through the population? However, this solution is short-sighted, as once the population has successfully responded to selection (i.e. reached a local fitness peak on the landscape), asexual reproduction should once again be favored and rapidly outcompete sexually reproducing individuals. Thanks for a wide array of theoretical exploration of this problem, we now understand that sexual recombination is much more likely to be favored in fluctuating environments, where populations must continually respond to changing selection pressures. Among the most consistently fluctuating selection pressures are those imposed by coevolving parasites and pathogens. This idea has been coined the “Red Queen hypothesis for the Evolution of Sexual Reproduction.” In the case of pathogens/parasites that are adapting to their host population, we expect that any mutation that allows pathogens/parasites to be more infective (successful) on a common host genotype to have a selective advantage and thus increase in frequency over time. However, this leads to a reduction in that common host over time (as pathogens/parasites necessarily harm their hosts) and increases the relative fitness of rare genotypes that are not targeted by the pathogen/parasite population. This is termed negative frequency-dependent selection because rare host types always have an advantage, but only until they themselves become common. Study questions: Why would negative, but not positive, frequency-dependent selection be a likely driver of sexual reproduction? Can you explain the seemingly counter-intuitive finding from the New Zealand mud snail system that sexually reproducing populations tend to have MORE parasites than asexually reproducing ones and describe why this is in fact strong evidence FOR the Red Queen hypothesis? Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 12 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. Coevolution and Cospeciation Coevolution is a process of reciprocal evolutionary change in interacting species. Or, more precisely, it is the reciprocal genetic change in interacting species, owing to selection imposed by each on the other. It can occur between any two populations as a result of ecological interactions, but these need not be direct (e.g. the two organisms do not have to ‘meet’ but could still impact one another’s evolution). We expect coevolution to occur any time two organisms impact one another’s fitness often enough to act as meaningful selection pressures on each other’s populations AND if there is heritable genetic variation underlying the traits under selection. Coevolution is more likely when these interactions are specific, rather than general. Thought experiment: Imagine if and how coevolution might occur in each of the two scenarios – (i) there is a bee species that pollinates over 100 different plant species in the local environment, (ii) there is a bee species that pollinates one rare species of plant in the local environment. In each case, how might you expect the pollinator to adapt to the plant(s) and how might the plant adapt to the pollinator? Under which circumstance might you expect the evolution of each of the following flowers? As shown in the table below, there are many possible coevolutionary interactions and they can be usefully characterized based on how each species impacts the other’s fitness. In the case of plant-pollinator interactions, for example, we would consider these to be mutualistic interactions because the pollinator receives resources from the plant and the plant receives a mechanism for sexual reproduction through the movement of pollen. One way to study coevolution is to measure how well adapted each population is to the ‘sympatric’ local population of the other species. In the case of a mutualist, for example, we might expect that one population of bees is more effective at pollinating the local population of a plant species than a population from a different area (for example on the other side of a mountain range, where gene flow is expected to be restricted). In the case of host-parasite coevolution, we would say that a parasite is locally adapted if it is more infective to its local host population than it would be to an allopatric, foreign population. Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 13 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. Table of possible coevolutionary interactions, based on whether each population has a positive (+), neutral (0) or negative (-) fitness impact on the other. + o - Antagonism + Mutualism Müllerian mimicry Commensalism Parasitism Batesian mimicry Predation Plant-pollinator Herbivory O Commensalism X Spite Batesian mimicry - Antagonism Parasitism Spite Competition Predation Herbivory Just like any other form of evolution in response to selection, we expect coevolution to be more rapid when/where there is more genetic variation upon which selection can act. In the case of two coevolving species who are antagonists, such that adaptation of one population comes at the fitness cost of the other, we expect that the species with more dispersal should be the one to be locally adapted (i.e. winning the coevolutionary race). For example, if a host has more dispersal than its parasites, we predict the host will have more genetic variation (constantly introduced by migration) and thus will more effectively respond to parasite-mediated selection than the parasite can counter-adapt. In this case, we might expect the host to be locally adapted, such that hosts are more resistant to local parasites than they would be to allopatric/foreign parasites. For predator-prey interactions (which are also antagonistic), it can be useful to separate out the ecological dynamics from the evolutionary ones. In this case, when we are describing the ecology we are referring to impacts of one species on the population size and/or range distribution of the other. We expect that increased numbers of predators will decrease the population size of their prey – and then that this decreased prey Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 14 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. will lead to a decrease in the population density of the predator. Without any evolutionary change per se, we therefore might expect cycles of predator and prey population densities over time. What happens when we begin to incorporate (co)evolution into this mix? In this case, we might predict that predators in an environment where prey are scarce will be under stronger selection to adapt to their prey (i.e. evolve better predation strategies). This would lead to even fewer prey and this cycle could continue UNLESS the prey populations also respond to predator-mediated selection. In this case, when a prey species is in an environment with high predator densities, we expect the prey to be under very strong selection and to adapt to the predator (i.e. evolve better defense mechanisms to predation). Experimental evolution can be a very powerful way to test the impact of one species on the other and to monitor evolutionary changes. In the case of coevolution, researchers can do something particularly cool: they can separate evolution from coevolution and ask what happens. In this case, you could experimentally ‘allow’ one species to evolve to the other over time and then in one treatment allow the second species to respond to changes in the first and coevolve, while in the other treatment prevent the second species from evolving by continually replacing it with individual from the original (ancestral) population. In one such study, Steve Paterson and colleagues either evolved or coevolved phages with their bacterial hosts and found that the rate of molecular evolution in the phage was far higher when both bacterium and phage coevolved with each other than when phage evolved against a constant host genotype. Coevolution also resulted in far greater genetic divergence between replicate populations, which was correlated with the range of hosts that coevolved phage were able to infect. Study question: Why would you expect more mutations to occur over evolutionary time in the phages that are coevolving with their hosts compared to those that are evolving in response to a ‘fixed’ host population? While experimental evolution allows researchers to observe coevolution in real time, we can also look for signatures of coevolution at the macro-evolutionary scale. A cophylogeny (as seen on the left) is a statistical method known as phylogenetic reconciliation analysis that tests the hypothesis that two phylogenies are more similar/different than expected by chance. If two species are obligately associated with one another (such as an endosymbiotic bacteria inside of an insect species that is vertically transmitted from parent to offspring and confers a fitness advantage to those insects carrying it), we would likely expect the speciation events in the insect phylogeny to match those of the symbiont. Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 15 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. Of course, there are many reasons that two coevolving organisms would NOT have nicely aligned phylogenies as well, and while coevolution can result in cospeciation and cospeciation patterns are likely in the case of tightly coevolving organisms (like figs and fig wasps), cospeciation is NOT definitive evidence for coevolution. This is for two reasons: (i) it might be that two species that are not interacting were affected by the same speciation processes as one another (e.g. formation of dispersal barriers that affected both species); and (ii) it might be that one species is adapting to the other (but not driving any evolutionary response/acting as an important selection pressure) and thus is speciating as a result of the speciation events occurring in the second species. In this latter case, one of the species is tightly ‘tracking’ the other, but not affecting the other’s fitness in any meaningful way and thus not coevolving. Eco-Evolutionary Dynamics As described in the case of predator-prey dynamics discussed above, ecological and evolutionary change can happen on the same timescale and can/do impact one another. In that case, increasing population densities of predators (ecology) could change the selection pressure acting on the prey populations, leading to evolutionary change in the prey as they evolve new defense traits. This change then feeds back on the ecology as it then reduces the number of predators. This is described as an eco-evolutionary dynamic (and sometimes as an ecological feedback). For such dynamics to occur, there must be (i) a strong effect of the phenotype/trait on the environment such that the organism structures (or ‘constructs’) their environment, thus changing the ecology of the system; and (ii) the constructed environment must lead to selection that drives the subsequent evolution of the population in question. In other words, changes in the environment cause selection on the population (directional or disruptive), and the population has sufficient genetic capacity to evolve in response to changes in its environment. As an interesting set of examples, let’s consider the evolution of host resistance to parasites (mechanisms that reduce within-host growth of parasites/pathogens) compared to the evolution of host tolerance to parasites (mechanisms that reduce the harm caused to the host without reducing parasite/pathogen growth). In the case of resistance, we would expect that as alleles conferring resistance spread through a population, the density/prevalence of the parasite/pathogen should decrease as well (because there are less susceptible hosts around). As such, the strength of parasite-mediated selection on the host population should also decreased. If resistance is costly (which is considered the rule given evolutionary tradeoffs and finite resources) then resistance should be selected against as parasite/pathogen density decreases. This eco-evolutionary cycle will then continue indefinitely – selecting for resistance when its rare and selecting against it when common, all due to the ecology of the system. Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 16 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. Study question: in the example above, what might you expect to happen to a parasite with a counter-resistance mechanism that is also costly? In contrast to resistance, tolerance evolution does not negatively impact the parasite/pathogen population, and thus as alleles conferring host tolerance spread through a population, we expect the parasite/pathogen density to INCREASE, rather than decrease. As such, the parasite-mediated selection will continue to drive evolution of the host population towards the fixation of the tolerance alleles. From Selfish Genes (mobile genetic elements) to Altruism (Inclusive Fitness) Evolution happens at the scale of a population (i.e. individuals cannot evolve) and the unit of selection (i.e. how fitness is determined) is the individual. Some genotypes have more offspring than others – if that difference is stochastic (i.e. has nothing to do with fitness, but rather chance) that neutral evolution can occur, and if it is the result of differences in fitness of this genotype relative to others, than adaptive evolution can occur and the population can move up a fitness peak. Because selection acts on individuals, and individuals are the sum of their parts (AKA genes), we don’t typically consider the fitness of an individual gene. That makes sense in most cases, but for some genes – the so-called mobile genetic elements - they have an inherent fitness of their own, as they are capable of ‘reproducing ‘outside of normal reproduction. Remember that a gene is simply a sequence of nucleotides along the DNA strand that determines the sequence of amino acids in a protein. But in the case of selfish ‘genes’ we will broaden this definition to include non-coding DNA ‘elements’ capable of replicating themselves. Some ‘selfish genes’ can replicate themselves within a genome, such that while a parent might have one copy of the gene, their offspring might have two. Other ‘selfish genes’ are capable of leaving their own genome entirely and moving into another via something we call horizontal gene transfer (the movement of genetic material between organisms that are not related by parent and offspring). Bacterial evolution is shaped by rampant horizontal gene transfer, and there are mechanisms that bacteria have evolved to increase its frequency. For example, some bacteria are able to pick up DNA from the surrounding environment and integrate into their own genomes (transformation), while others have structures (pili) that act as a bridge between two cells and can transmit plasmids (small circular genomes that are not part of the core bacterial genome) between them (conjugation). Finally, many bacteriophages are capable of packaging host DNA within the viral capsids and then introducing them into the next host cell that they infect (transduction). Overall, the vast majority of these HGT events Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 17 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. will be – like all random mutations – detrimental and the host lineage containing the new genes/elements will be selected out of the population. But occasionally these horizontally transmitted genes will confer a fitness advantage to their new host and spread through the population. However, horizontal gene transfer is not just for the bacteria: in humans, approximately 50% of the genome are thought to be mobile genetic elements (MGEs). In plants and animals, the most common type of MGEs are transposons (or transposable elements; TEs), which can move between DNA molecules or within a single DNA molecule. TEs are ubiquitous throughout eukaryotic genomes and have been shown to alter regulatory networks, gene expression, and to rearrange genomes as a result of their transposition. As they jump in and out of genomes, they can disrupt genes (e.g. by landing in the middle of them!) and thus rapidly alter the organisms’ phenotype. One super cool fact about TEs is that some of them can ‘mobilize’ when their host organism experiences stress. Study question: why might it be evolutionary advantageous for a TE to jump around the genome in times of stress? Other MGEs/Selfish genes include endogenous retroviruses (viral elements in the genome that closely resemble and can be derived from retroviruses, those viruses that must integrate into and then excise from host genomes to complete their replication/life cycle), and segregation distorter genes (a class of genes that emerge from meiosis in more than one-half of the gametes produced by a heterozygote). The other reason it can be useful to think about evolution through the lens of individual genes is because one individual who ‘helps’ another (i.e. contributes to the increased fitness of another individual) at the expense of themselves can actually realize a benefit as a result of something we call kin selection. Here, even though the focal individual isn’t passing on their own genes directly, if they help a related individual, they can increase the likelihood that their genes make it into the next generation via the reproduction of their kin. Inclusive fitness is the sum of an organism's direct fitness (i.e. their offspring) and indirect fitness (the number of offspring its relatives have, multiplied by the degree of relatedness). The idea of inclusive fitness became particularly important as a means to explain the mystery of altruism (behavior by an individual that increases the fitness of another individual while decreasing their own) in nature. Hopefully by now it is clear why a gene encoding a trait that benefits another individual at the direct cost to the holder of the gene would be unlikely to spread by natural selection. However, altruism exists in nature and thus need an explanation. Hamilton's rule puts forward the idea that an altruistic behavior can spread in a population when br-c> 0, where c is the cost to the altruist, b is the benefit to the recipient, and r is the coefficient of genealogical relatedness, which ranges from 0 (for unrelated Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 18 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. individuals) to 1 (for identical twins). When this inequality is satisfied, an altruistic act increases the absolute number of copies of the altruistic gene. When r = 0 (no relatedness), a gene encoding altruism is not predicted to spread by positive selection, and should be selected against if there is any cost at all. Another clear solution to the ‘problem’ of altruism is that individuals interact with other individuals more than once, so you have more information about whether they tend to cooperate or cheat. However, this requires that you can recognize individuals as ‘cheaters’ or ‘cooperators. A 'green-beard' gene is defined as a gene that causes a phenotypic effect (such as the presence of a green beard or any other conspicuous feature that can be detected/recognized by other individuals) that allows the bearer of this feature to recognize it in other individuals, and causes the bearer to behave differently towards other individuals depending on whether or not they possess the feature. If there is a conspicuous trait (i.e. a greenbeard) that is only present in those individuals that cooperate, for example, than cooperators can specifically cooperate only with other cooperators and therefore reduce the fitness of cheats in the population. Study question: What would happen if a greenbeard gene occurred on one chromosome and the gene encoding cooperation was encoded on another chromosome if the organism was a sexually reproducing one? However, the problem of altruism can also be solved by thinking about Group selection occurs when traits evolve according to the fitness (survival and reproductive success) of groups, not individuals. This is considered a relatively rare from of natural selection, but can occur any time that a trait has a particular group context (imagine a herd of sheep or pack of wolves) such that groups with many individuals carrying this trait are more/less fit (as a group) than other groups with fewer individuals carrying that trait. Mathematically, group selection occurs when the overall group fitness is higher or lower than the mean of the individual members' fitness values. Sexual Selection and Mate Choice When resources are scarce in the environment, there will be selection for increased competitive ability among those individuals that rely on that resource for reproduction. In the case of sexual selection, the ‘resource’ under competition is mates! And thus any trait that increases one individual’s likelihood of successfully finding a mate and reproducing should be favored by selection and spread in the population through time. Mate competition occurs when mates are a limiting resource, and thus when sexual selection is likely to shape fitness (intrasexual competition). Mate choice occurs when there is a fitness advantage to selecting particular mate phenotypes, and in this case sexual selection shapes who is likely to be selected as a mate (intersexual interactions). Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 19 Copyright © 2024 Britt Koskella, all rights reserved. Do not copy, distribute or upload to any website. As a general (but not exhaustive!) rule, the non-offspring bearing sex is often under higher mate competition than the offspring-bearing sex. This is because, while the fitness of offspring-bearing individuals is limited by how many offspring they can physically produce, the fitness of non-offspring-bearing individuals is shaped by how many times they can mate (and that can be very large!). This is why we tend to see dramatic sexual dimorphism in many species, where one sex looks or behaves very different than the other. This typically tells us that the trait that is dimorphic is likely to be under sexual selection. Importantly, if there is large variation in reproductive success among individuals as a result of unequal access to mates, this will lead to strong sexual selection. If all individuals tend to have the same number of mates, then we expect very weak, or no, sexual selection to occur. In contrast to mate competition, mate choice results from individuals being choosy about the traits their potential mates carry. If a trait (e.g. the showy tail of a peacock) is an honest signal about general health, then any individual who can differentiate mates based on this trait and choose the ‘fitter’ mate will have ‘fitter’ offspring. A showy tail is a costly trait, and it may well be that to grow and maintain such a tail one must be well fed, resistant to pathogens and disease, and able to evade predation. In this case, choosing a mate with a very showy tail means you are more likely to have offspring that carry all the other alleles that underly that high fitness in the current environment. Runaway sexual selection is occurs when the trait under sexual selection becomes more and more extreme as a result of mate choice, even to the point of becoming too costly to maintain under natural selection (e.g. at some point the benefit received from additional matings is outweighed by other costs, such as predation). Runaway selection assumes that sexual preference in females and ornamentation in males are both genetically variable and likely linked, such that sons from ‘choosy’ mothers are more likely to ornamented and daughters from ornamented fathers are more likely to be ‘choosy.’ As each trait (choosiness and ornamentation) are selected for, they reinforce one another and continue to escalate. Core Concepts/Study Guide from Koskella lectures Fall 2024 IB 160: Evolution 20