Sunday, 25 November 2012

Part 2: On testing the Red Queen hypothesis

The Red Queen Hypothesis
A Hamiltonian parasite is a pathogen or parasite whose variation-selection cycle is clearly shorter than that of its host (see here and here or Hamilton 1980, 1982, 1990). In order to compensate this asymmetry in variation-selection cycles, the host may boost its rate of variation through sexual recombination. At the same time, asexual (clonal) hosts may be particularly vulnerable to Hamiltonian parasites, because they tend to crop up in genotype mono-cultures. The Red Queen hypothesis suggests that Hamiltonian parasites exterminate clonal but not sexual hosts and maintain sexual reproduction in the face of asexual competitors that way.

Host specificity
If parasites were not host-specific, however, selection should not favor such a specialization on common genotypes, that is, asexual hosts.
    Suppose a parasite has to be able to infest individuals of several alternative host species, in order to survive. Any host species might simply not be enough to sustain a parasite lineage. If such a parasite nevertheless specializes on a common genotype within any one of its host species, this will probably mean its imminent extinction. Mutations that make this parasite more generalist (and less specific to particular host genotype) might therefore survive in the parasite gene pool longer than parasites that specialize. Let's call it the rare-host effect. The arms race suggested by the Red Queen hypothesis might not take place, if the parasite is not host-specific.
    If an asymmetry in the variation-selection cycles favors a parasite that is host-specific, then the parasite will adapt to common genotypes of its host. It may therefore affect asexual (clonal) hosts worse than sexual ones, because the former tend to crop up as genotype mono-cultures.

Asymmetries in selection pressure
Dawkins (1982, chap. 4) explains two further asymmetries in selection pressures. The cases he considers are without a marked asymmetry in the variation-selection cycle of the interacting species, for example, insect-parasitoid or predator-prey interactions. Nevertheless, asymmetries in selection pressure are possible.
   A parasite may be rare and therefore not exert a large selection pressure on the host (see rare-enemy effect in Dawkins 1982). Conversely, the fitness reduction of losing a round of interaction might mean a loss of a dinner for the enemy but a loss of life for the prey (see life/dinner principle in Dawkins 1982). In host-parasite systems the life/dinner principle often works out the other way round. In a round of interaction the parasite will lose its life, if it fails to infest a potential host, whereas the host will only lose a dinner, if the parasite's effect on its fitness is mild. A parasite with a strong negative effect on the host's fitness will produce a symmetric situation.
    From Dawkins (1982) we can conclude that the arms race will be called off, if one side wins due to an asymmetry clearly favoring it. An arms race will only ensue if no side is a clear winner, that is, if the asymmetries in the system roughly cancel each other.
    For example, let the variation-selection cycle of a parasite be half as long as that of its host and the risk of a host to contract this parasite be 0.5. The rare-enemy effect will cancel the variation-selection asymmetry and an arms race will ensue. There will then, however, remain no need for sexual recombination to cancel the variation-selection asymmetry. Hence the Red Queen hypothesis further needs a parasite who is common (no rare-enemy effect) and has a strongly negative effect on the fitness of the host (no life/dinner effect). This is often noted as restrictive requirements for the parasite in modeling papers.

An arms race will be on, when asymmetries in variation, selection, or the variation-selection cycle roughly cancel each other. Whenever sexual recombination keeps the host in an arms race with its parasite, asexual hosts should drop out of that race (the Red Queen hypothesis will hold). Actually an arms race could even be on between a sexual parasitoid and its sexual host, and asexual host mutants would drop out of the race. That is, my previous post would be wrong in thinking that parasitoids do not qualify for Red Queen dynamics in principle.Nevertheless, there should  be many combinations in which the asymmetries do not balance and the arms race will be called off to one side being a clear winner. That is enough scope for reasons why half the systems studied so far yielded negative evidence on the Red Queen (see previous post).
  • Dawkins, R. (1982) The extended phenotype. Oxford University Press.