NOTE: THE INFORMATION ON THIS PAGE IS NO LONGER PART OF THE COURSE (removed from main wiki, 2014-2015)

Prev: Multiple coding
Next: Integrating pattern formation, coding structure and evolution


TODO LIST
  • Instead we just have an idea of the dynamical system such operators define the outcome of which is not trivial in the sense of going somewhere (NOT CLEAR), i.e the dynamical system gives some structure ... what is it? (continued later)???
  • REF: that cross-over might be sensible for fast evolution such as co-evolution (Hamilton REF).
  • REF: studied RNA strings co-evolution in space with cross-over



Breaking down the landscape


We have already mentioned the short comings of the landscape metaphor. Here we focus on the nearness concept that a landscape metaphor requires. Obviously this is clear with point mutations, but becomes increasingly undefined with other mutational operators. Operators such as cross-over, are highly non-local, how should be consider them? What is their effect on mutational dynamics, attractors, domains of attractors and side-effects? Well, we do not have a clear a priori idea about such dynamics. Instead we just have an idea of the dynamical system such operators define the outcome of which is not trivial in the sense of going somewhere (NOT CLEAR), i.e the dynamical system gives some structure ... what is it?
(continued later)

The other weak aspect of the landscape metaphor is its static nature, i.e. fixed in time, whereby evolution is assumed to occur on a very long timescale. In line with this is that most evolutionary simulations appear to take millions of generations, while in biology in such timescales amazing changes occur: climate change, geological periods, speciation. Simulations based on a fixed fitness landscape metaphor therefore appear to be out of sync with biology in this sense.

However not everything changes. The newest mammal, the polar bear, is very recent, but also several mammals have remained very much the same. While when we think of a co-evolutionary processes (e.g. host-pathogen systems) timescales must be very short! Other processes however may not depend on the external environment that much, e.g. intrinsic processes. However, it is clear that for most cases the fixed fitness landscape cannot be right.

With respect to the RNA world we do however see an intermediate concept: the primary structure to secondary structure mapping, the latter of which may define a varying fitness. Although the fixed fitness concept does not apply, the RNA structure / energy landscape does apply as fixed landscape. This in itself is nice in order to know how conclusions based on the landscape apply when the landscape is not there.

RNA co-evolution

Here we study the effects of co-evolution in the RNA world on coding structure (Huynen & Hogeweg 1994). Simulations are set-up not with secondary structure defining fitness, but two population of RNAs. One in the host which does not want a match with the parasite. The parasite needs to match the host for fitness.

In this setting one obtains the Red Queen dynamics of co-evolution, and in contrast to evolving to fixed structures, here there is evolution to non-robustness! Why does this happen?
  • the host just wants to change as much as possible to be different and so mutations should have a large impact
  • the parasite needs to follow, but can't be as non-robust as the host because it needs to match somewhat (need information)
  • this is why neutrality decreases!

What does this mean for the biological context?
  • robustness comes for free if the optimum does not change in time
  • with the need to change one get the oposite coding: non-neutrality
  • both can happen in the same genome!

Co-evolution and cross-over: sexual and asexual "prey" in RNA

In the Royal Road we saw that cross-over doesn't help, so why then cross-over (or sex)? This is a much debated issue which is not very much resolved: i.e. why add foreign genome. One thing is clear however, that cross-over might be sensible for fast evolution such as co-evolution (Hamilton REF). (REF) studied RNA strings co-evolution in space with cross-over. Results show that sex at least stays in the population, but only in the co-evolution context and not with fixed external situation.

Next: Integrating pattern formation, coding structure and evolution


References
Huynen MA & P Hogeweg (1994) Pattern generation in molecular evolution: exploitation of the variation in RNA landscapes. J. Mol. Evol. 39, 71-79.