Meet the bloggers:
Mitchel Daniel is a postdoctoral fellow at Florida State University. He is an evolutionary and behavioral ecologist, and is especially interested in sexual selection, kin selection, and kin recognition. Follow Mitchel’s work @MitchelJDaniel.
Walid Mawass is an evolutionary geneticist, currently a PhD candidate at the Université du Québec à Trois-Rivières, studying the evolution of life histories in French-Canadian historical populations using quantitative genetic and genomic approaches. Walid is interested in evolutionary genetics theory in general, with a current focus on contemporary evolution in natural populations and the role of interactions between environment and genetics on evolutionary trajectories. Follow Walid’s work @MawassWalid.
Sabrina Heiser is a PhD Candidate in Dr. Charles D. Amsler’s lab at the University of Alabama at Birmingham. Her research focuses on the factors driving the geographic distribution of chemical defenses in a red seaweed. For her sample and data collection, she gets to go and SCUBA dive in Antarctica. She received her B.Sc. in Marine Biology from Plymouth University (UK) and is originally from Germany. Follow her adventures on her website.
Weeds are responsible for massive agricultural costs associated with herbicide usage and lost productivity. Feral weedy rice has evolved repeatedly and is highly competitive, reducing the yield of rice fields by up to a whopping 80%. To better understand how we might deal with this weed, Marshall looked for evidence of introgression between weedy rice and co-occurring wild and cultivated populations in Thailand. Examination of neutral regions suggests that weedy strains match the local cultivars, implying introgression. Discordant evidence from domestication genes suggests that recombination has made these two signals no longer congruent. Marshall underscored some ominous implications of his work. It is “disturbingly easy to evolve weedy rice,” and there are conservation concerns related to the escape of transgenes into wild ancestors and weedy populations, which mean that many herbicides will only be effective for so long.
How can we incorporate fitness interactions into interacting phenotypes theory? Joel is helping to address this question by blending interacting phenotypes theory and game theory. These two frameworks are two sides of the same coin – the former focusing on evolutionary processes and the latter on the products of evolution. These frameworks have frequently been used to study similar biological problems, but have historically been viewed as alternatives. Blending them seems like a challenging but exciting and powerful approach for understanding social evolution. One novel implication of Joel’s work is that IGEs can help cooperative behaviors with synergistic effects to invade a population, which would otherwise be difficult.
Covariances among traits can lead to correlational selection, altering net selection on those traits. Trait covariances are, of course, affected by patterns of assortment, but how can non-random assortment arise? Butch is using forked fungus beetles to help address this question. The habitat of these beetles – the fruiting bodies of wood-decaying fungi – are patchy, and males wrestle with their neighbors using their horns. Size is important for winning these contests, it’s good to be a big male surrounded by small males. Looking at the metapopulation level, covariation in size is close to 0; but, at the subpopulation level, covariation varies from strongly negative to slightly positive. Clever experiments helped to shed light on why. Environments with dispersed resources (fungi) had more negative assortment. Groups consisting of highly social individuals tended to exhibit positive assortment, while groups with less social individuals did not. Butch interprets these patterns as suggesting that more interaction leads to less negative assortment. Though the exact social mechanism(s) involved are still a bit of a mystery, it’s clear that patterns of assortment are highly context-dependent, and so vary among subpopulations. This is interesting because it creates the potential for multilevel social selection. The work also prompted lively discussion about how best to assess social context, including what spatial scales are relevant – a classic problem that remains challenging to grapple with.
As an aside, forked fungus beetles seem like such a great study organism for exploring the intersection between social behavior, spatial ecology, and evolution. I’m always stoked to see the newest data coming out of this system.
Microbes are many, and they often live in close symbiosis with hosts. Microbes can affect many host traits, with consequences for both host and microbial fitness. When microbes and their hosts differ in the optimal value of these shared traits, conflict can arise. It was exciting to follow Maren’s theory exploring the evolutionary dynamics that ensue. To summarize her findings, fitness conflict leads to sustained compensatory coevolution (a ‘tug-of-war’ over the trait value). One prediction is that this dynamic leads to the evolution of codependence, with either party being maladapted in the absence of the other. Maren’s model also predicts that fitness feedbacks will constrain trait conflict, and that conflict over these multi-genomic traits can maintain high levels of allelic variation in both partners.
Joel Mcglothlin – Interacting phenotypes and interacting fitness
Correlational selection is a way that traits interact with each other. Positive correlational selection can lead to stronger genetic association between traits and that can be represented by a strong positive genetic correlation. Jump to the theory of interacting phenotypes, which is defined when two individuals interact with each other (the famous psi), and this implies indirect genetic effects. In this case as well, we can have both non-social selection and social selection on fitness. In the simple model, there are no fitness interactions (meaning all paths of effects are linear). This raised the question: is it possible to incorporate fitness interactions into interacting phenotypes theory? And do social fitness interactions influence genetic correlations? One way is to draw parallels between the interacting phenotypes (IGEs, concerned with how they affect response to selection) and evolutionary game theory (genetic variation is assumed and we are concerned with the outcomes of selection). A staple of game theory is the pay-off matrix giving fitness effects from different interactions between two partners (traits expressed as binary 0/1, i.e. two different strategies possible). Using some symbolic substitution, we can arrive at a fitness equation to compare with the classic fitness equation. Between the two, we find that the synergistic effect of the interaction plays an important role. With some mathematical operations, we find that both the social and non-social selection gradients are (global) frequency-dependent because they depend on the mean of the phenotype in the total population. The social selection gradient can be considered local frequency-dependent because it depends on the local partners that the focal individuals interact with. Next, these results can be applied to any game theory model, e.g. the prisoner’s dilemma. Based on modelling results, a weak synergistic effect leads to a constantly negative non-social selection gradient. However, when this synergistic effect is increased, then both the social and non-social selection gradients will increase. This can be extended to Hamilton’s rule where we can add the effect of relatedness and IGEs. Adding the synergy effect will add more weight to the cooperation part of Hamilton’s rule. So IGEs can help cooperative behaviors with synergistic fitness effects invade a population.
Estimating social fitness interactions, involves adding the usual Lande-Arnold regression question non-social/social interaction selection gradients (non-linear). So can these selections gradients lead to changes in genetic correlations like correlational selection can. These might play a role in cases where we have displays, either aggressive or courtship, i.e. signalling.
Butch Brodie – Phenotypic assortment changes the landscape of selection
The social context is important to evolutionary processes. It is not only social animals that act socially, but any organism that interacts with a conspecific. Social context can impact fitness, though social selection, through phenotypic expression, through psi and IGEs, and finally the targets and models of selection. In a multivariate framework of social selection, we have both the nonsocial and social selection on the individual fitness. The focus here is the interactant covariance between the focal individual’s trait and the average phenotype of the partner. This can be extended to non-linear selection terms as well. The fitness effects of social selection are general. But the phenotypic effects of social selection are trait specific. The net effect of the social selection can change with the interactant covariance. In the case of random assortment, the interactant covariance is null. Hence, the social selection will be distributed randomly among phenotypes, so no effects of social selection. However, if we have positive assortment (like forms assort together), the direction of social selection on a given trait is reinforced. In another case, we can have negative assortment (unlike types assortment), so the effect of the social selection gradient is reversed. Sources of the interactant covariance can arise from active non-random assortment (behavioral or altruistic kin associations), passive non-random assortment (environmental choices, inbreeding, spatial structure), phenotypic modification where the phenotypes are changed through IGEs or common environmental effects, and, finally, social group size. Average relatedness can generate positive interactant covariance across the phenotype which can reinforce the direction of social selection. However the caveat here is that it might reinforce as well apterns of genetic covariances that can impact social selection. Behavioral assortment or modiviation can lead to trait specific associations which can have unpredictable effects on social selection. Small groups tend to generate negative interactant covariance which can reverse the effects of social selection. What affects variation in interactant covariance?
Marshall Wedger, the winner of this year’s SJOB student paper prize, presented his work on weedy rice populations in Thailand. Humankind is constantly increasing its land use through agriculture to accommodate population growth. Crops are being artificially selected to maximize output through increasing yield and resistance towards certain conditions, pests or diseases. Something I have less thought about is Wedger’s work on introgression of weedy crops into cultivated populations. His population genetic study showed that weedy rice indeed hybridizes with cultivated populations in Thailand which can have negative implications on the harvest. It does appear, that the presence of wild rice populations facilitated this process as the same was not observed in US crops of rice.
Maren Friesen looked at how host-microbe interactions impact the phenotypes and fitness of the host. She used Arabidopsis and Pseudomonas as the model system and looks at the root branching as a trait. Her work found that traits were often dependent on the genome of both partners which are termed multi-genomic traits. Evolution takes place in both partners, contributing to the overall phenotype. Last but not least, variation as well as coevolution is maintained through conflicts in fitness of those multi-genomic traits.