Selection mosaics and the GMTC

This past week in Coevolvers, we dropped back into the empirical world and ready a paper from Piculell et al (2008) on evidence of selection mosaics. Selection mosaics describe a case where the fitness function of the interacting players varies across space (Gomulkiewicz et al 2007; Thompson 1999, 2005), sometimes described as GxGxE interactions (G: genetic; E: environment). What does this mean more generally? Simply put, the fitness of a plant may change from one population to the next because the nature of the interaction with a mutualist is affected by the environment. This can occur even if the genotypes that make up those populations are exactly the same.

The experimental design was certainly setting up the case for a maximum chance of detection of interaction effects. With only levels of each factor, (e.g. two genotypes of the host) the authors had less power to detect any main effects, but that clearly wasn't the objective. They wanted to find evidence of significant GxGxE. Essentially this experiment had 4 environmental treatments, so they maximized the chance of an interaction. The authors of this paper were very upfront that they were not intending to measure a selection mosaic in the natural setting. Their objective was to demonstrate the possibility and they certainly obtained that goal. With that limitation in mind, how general are these results? Measuring the potential for a selection mosaic is one thing, but for this to really have an impact in generating or maintaining diversity as imagined in the Geographic Mosaic Theory of Coevolution (Thompson 1999, 2005) then it must hold for a broad sample of the populations under investigation. The authors are on a good track though to discovering more about this system. Perhaps they plan on taking the methodology outlined in Nuismer and Gandon (2008) on reciprocal-transplant designs. Picking a larger sample of the genetic variation found in nature for at least one of the players would extend their results from the possible into the probable.

References

Gomulkiewicz, R., D. M. Drown, M. F. Dybdahl, W. Godsoe, S. L. Nuismer, K. M. Pepin, B. J. Ridenhour, C. I. Smith, and J. B. Yoder. 2007. Dos and don'ts of testing the geographic mosaic theory of coevolution. Heredity 98:249-258.

Nuismer, S. L., and S. Gandon. 2008. Moving beyond Common-Garden and Transplant Designs: Insight into the Causes of Local Adaptation in Species Interactions. American Naturalist 171:658-668.

Piculell, B., J. Hoeksema, and J. Thompson. 2008. Interactions of biotic and abiotic environmental factors in an ectomycorrhizal symbiosis, and the potential for selection mosaics. Bmc Biol 6:23.

Thompson, J. N. 1999. Specific hypotheses on the geographic mosaic of coevolution. American Naturalist 153:S1-S14.

Thompson, J. N. 2005.
The Geographic Mosaic of Coevolution. University of Chicago Press, Chicago.

Paper Read

Piculell, B., Hoeksema, J., & Thompson, J. (2008). Interactions of biotic and abiotic environmental factors on an ectomycorrhizal symbiosis, and the potential for selection mosaics BMC Biology, 6 (1) DOI: 10.1186/1741-7007-6-23

Does nestedness lead to more nestedness?

This past week in Coevolvers, we read a brand new paper in Nature from Bastolla et al (2009). The authors create a simple model to understand how network structure can lead to an increase in predicted biodiversity in a community. In this case, the authors were looking at how a network of mutualistic interactions will generally be nested. This network structure can reduce interspecific competition and allow a greater biodiversity. The nestedness of interactions in this kind of community refers to how many pollinators a pair of plants share compared to their total number of pollinators. The more they share, and the more this is the case across the entire network, then the higher the network nestedness. The authors use a set of previously published real networks to test predictions from their model.

I thought I would have a quick look at some of these "real" networks. The appendix of the paper directed me to Bascompte et al (2003). This paper summarized pollinator, seed dispersal, and food web networks of plant-animal interactions.

While there I noticed a reference to a review paper in Annals of Botany (Vazquez et al 2009) with an exciting title (Uniting pattern and process in plant-animal mutualistic networks). This looks like a great review and perhaps a future post. In the section outlining "patterns", they provide two contrasting topics, "Mutualistic networks tend to nested" but also "Mutualistic networks tend to be compartmentalized". This struck me as contradictory to the paper we read (Bastolla et al 2009) which predicted nested networks to emerge.

Vazquez et al (2009) had several citations for compartmentalized networks (Dicks et al 2002; Guimaraes et al 2007; Olesen et al 2007). I looked up the Dicks et al paper. I see they find compartmentalization. "The compartments reflected classic pollination syndromes to some extent, dividing the insect fauna into a group of butterflies and bees, and a group of flies, at both sites. The compartmentalization was also affected by phenology" (Dicks et al 2002). There are certainly more examples out in nature that are compartmentalized. Pollinator syndromes could create these compartments. There are other examples of real networks of mutualisms that show compartmentalization. Vazquez et al (2009) finally point to a paper from Lewinsohn et al (2006) where they propose how both patterns can co-occur (compartments with nestedness) and I think this is really what Dicks et al (2002) is finding. Olesen et al (2007) have a paper where they are essentially calling this modularity. You have compartments (modules) and then nested networks present within those. While the original paper we read (Bastolla et al 2009) contained a potential process for how mutualistic networks can form, it seems as though natural networks are probably the result of a complex set of processes.

References

Bascompte, J., P. Jordano, C. J. Melian, and J. M. Olesen. 2003. The nested assembly of plant-animal mutualistic networks. Proceedings of the National Academy of Sciences of the United States of America 100:9383-9387.

Vazquez, D. P., N. Bluthgen, L. Cagnolo, and N. P. Chacoff. 2009. Uniting pattern and process in plant-animal mutualistic networks: a review. Ann Bot.

Dicks, L. V., S. A. Corbet, and R. F. Pywell. 2002. Compartmentalization in plant-insect flower visitor webs. Journal of Animal Ecology 71:32-43.

Olesen, J. M., J. Bascompte, Y. L. Dupont, and P. Jordano. 2007. The modularity of pollination networks. Proceedings of the National Academy of Sciences 104:19891-19896.

Lewinsohn, T., P. Prado, P. Jordano, J. Bascompte, and J. Olesen. 2006. Structure in plant-animal interaction assemblages. Oikos 113:174-184.

Paper Read

Bastolla, U., Fortuna, M., Pascual-García, A., Ferrera, A., Luque, B., & Bascompte, J. (2009). The architecture of mutualistic networks minimizes competition and increases biodiversity Nature, 458 (7241), 1018-1020 DOI: 10.1038/nature07950

Universal understanding of host-parasite adaptation

We recently read a theory paper by Gandon and Day (2009). In this paper they describe a valuable method for dissecting how interactions between a host and parasite alter mean fitness. Their method uses an understanding built from Fisher's fundamental theorem. They partition changes in mean fitness based on three different factors: natural selection, environmental change, and mutation. We know that the rate of adaptation is going to result from the amount of genetic variance in the focal organism (Fisher's theorem), but what about the impact of an interacting species that evolves as well (i.e. a coevolving parasite? Here is the real beauty of their analysis because the coevolving species becomes the environment. By separating the changes in a population mean fitness into changes driven by different forces, the authors provide not only a mathematically useful model, but also a useful intuition for understanding how hosts and parasites coevolve.

There are several ways that theoreticians often describe a host-parasite interaction (e.g. gene-for-gene, matching alleles) and these describe natural systems to some degree of accuracy. The authors use their method to analyze some recent empirical evidence (Buckling and Rainey 2002; Decaestecker et al 2007). They use the time series data on the interaction to test hypotheses of the nature of the interaction. These empirical studies compare the fitness of parasites against hosts from the past that they have coevolved with and those from the future (hosts that evolve later in the study). By making these comparisons, they have the ability to hold other factors constant (the genetic variance of the parasite population) and vary the environment (the hosts). Their model makes different predictions that should be evident from empirical evidence about how parasite mean fitness should change when the environment is varied.

The authors very elegant proposed method of looking at changes over time works well for systems where archives of past populations are possible as in experimental evolution systems (Buckling and Rainey 2002) or clever natural systems (Decaestecker et al 2007), but what about the rest of us? Addressed in at the very end, but only briefly, is a comparison of spatial patterns of coevolution when temporal data is missing. I think this issue of substituting space for time is potentially very powerful, but also somewhat more complicated. Temporal samples of a coevolutionary system could be predicted to have a certain amount of autocorrelation, but does this hold for spatially distributed systems. It certainly would nice to assume that there is a relationship between distance and time and this will of course depend on gene flow. How would selection mosaics (Gomulkiewicz et al 2007; Thompson 1999, 2005) impact this potential relationship? I look forward to future research as it provides some answers.

References

Buckling, A., and P. B. Rainey. 2002. Antagonistic coevolution between a bacterium and a bacteriophage. P Roy Soc Lond B Bio 269:931-936.

Decaestecker, E., S. Gaba, J. A. M. Raeymaekers, R. Stoks, L. Van Kerckhoven, D. Ebert, and L. De Meester. 2007. Host-parasite 'Red Queen' dynamics archived in pond sediment. Nature 450:870-873.

Gandon, S., and T. Day. 2009. Evolutionary epidemiology and the dynamics of adaptation. Evolution 63:826-838.

Gomulkiewicz, R., D. M. Drown, M. F. Dybdahl, W. Godsoe, S. L. Nuismer, K. M. Pepin, B. J. Ridenhour, C. I. Smith, and J. B. Yoder. 2007. Dos and don'ts of testing the geographic mosaic theory of coevolution. Heredity 98:249-258.

Thompson, J. N. 1999. Specific hypotheses on the geographic mosaic of coevolution. American Naturalist 153:S1-S14.

Thompson, J. N. 2005.
The Geographic Mosaic of Coevolution. University of Chicago Press, Chicago.

Paper read

Gandon, S., & Day, T. (2009). EVOLUTIONARY EPIDEMIOLOGY AND THE DYNAMICS OF ADAPTATION Evolution, 63 (4), 826-838 DOI: 10.1111/j.1558-5646.2009.00609.x

Where did this infection come from? Covert infections selected by demographic variability

This week we continued along our current path of pathogen models and looked at a recent paper (Sorrell et al 2009) investigating covert infections, a common and unexplained phenomenon of some pathogens exhibiting long periods of infection where they are silent (or covert in the language of the paper). During this silent/covert stage, the infections are mostly avirulent and non-infectious. These authors extend a previous SI type model that incorporated a covert state (Boots et al 2003) to understand what forces select for this kind of pathogen.

Extending a previous SI model (Boots et al 2003), the authors build a two strain model that includes susceptible hosts and multiple classes of infected hosts. With two strains, there are two broad types of infected hosts. Each of these is split again. The hosts can carry a covert infection or an overt infection. Covert infections are allowed to become overt but not the other way around. There are multiple trade-offs built into this model. A covert infection does not cause an increased host death rate (avirulent), but it does impose a cost to host fecundity where as an overt infection is virulent but does not decrease fecundity. In addition, covert infections are only transmitted vertically (from parent to offspring), while on the other hand overt infections are only transmitted horizontally (among individuals within the population).

Without additional forces, they find no selection for covert infections. However, given the abundance of such pathogens in nature, there must be some forces that are generating the proper conditions. The paper explores three different mechanisms that may be involved in selection for covert infections. The first examines the effect of superinfection (multiple pathogen strains in the same host). They conclude that selection will favor covert infections that are protective, that is they prevent superinfection. The other two mechanisms consider nonequilibrium host dynamics, temporal variation in host density and transmission. When variation is high and the potential to be lost from the population because of a lack of hosts or a lack of transmission events, then covert infections which again are vertically transmitted become more likely.

A question that was brought up during our discussion was: are these results different from a horizontal vertical transmission trade-off? When transmission opportunities are likely (high populations), then horizontally transmitting virulent pathogens are favored. In situations when there are fewer opportunities (e.g. during host population declines), then a pathogen that retains some vertical transmission and will be favored. Favoring a more covert pathogen is really just selecting for these two fixed trade-offs. I think what this paper contributes thought is a more thorough mechanistic explanation for how this trade-off works. They provide many biological examples of pathogens with complex covert behavior and this study certainly provides evidence of how they may have arisen.

This paper was quite interesting to me in that it was the first adaptive dynamics analysis that I've really understood. The authors walk through their methods and explain how to read the pairwise invisibility plots (PIPs) and provide some helpful but uncomplicated simulations too. Recently Dercole and Rinaldi (2008) published an introduction to this modeling/analysis technique that I'm looking forward to reading in the near future.

References

Boots, M., J. Greenman, D. Ross, R. Norman, R. Hails, and S. Sait. 2003. The population dynamical implications of covert infections in host-microparasite interactions. Journal of Animal Ecology 72:1064-1072.

Dercole, F., and S. Rinaldi. 2008. Analysis of Evolutionary Processes: The Adaptive Dynamics Approach and its Applications. Princeton University Press, Princeton.

Sorrell, I., A. White, A. B. Pedersen, R. S. Hails, and M. Boots. 2009. The evolution of covert, silent infection as a parasite strategy. Proceedings of the Royal Society B: Biological Sciences: online early.

Paper read

Sorrell, I., White, A., Pedersen, A., Hails, R., & Boots, M. (2009). The evolution of covert, silent infection as a parasite strategy Proceedings of the Royal Society B: Biological Sciences DOI: 10.1098/rspb.2008.1915