Predicting Evolution: A Beginner's Model

Silance writes "According to ScienceDaily , Scientists have developed a method of accelerating evolution in the lab that accurately mimics natural evolution. Drug-resistant E.coli strains from the 1940's that were subjected to the evolutionary speed-up process indeed followed the same evolutionary path as their natural bretheren. It is believed that the process could be used to predict the future monkey-wrenches that evolution might lob our way. Neat-o!"

Comments

Yet another sorry day for creationists.

[cp99]

With every paper like this one, the case for evolution gets stronger (not that it needed it), whereas the pesudoscientists falls apart (not that it hasn't already).

Artificial societies

[babbage]

I seriously doubt that any complex system such as evolution could ever be simulated with perfect accuracy. More likely, we can get to the point where likely outcomes can be guessed but the paths that lead to them will probably always be a surprise to us.

Check out the article on artificial societies from the current (April 2002) issue of Atlantic Monthly. I was thinking of submitting it to Slashdot anyway, but it particularly relates to this discussion too. The header blurb is:

The new science of artificial societies suggests that real ones are both more predictable and more surprising than we thought. Growing long-vanished civilizations and modern-day genocides on computers will probably never enable us to foresee the future in detail--but we might learn to anticipate the kinds of events that lie ahead, and where to look for interventions that might work

The article goes on to discuss many applications of this technique. None of them are specifically about genetic evolution, though one does analyze the settlement patterns of a pre-Columbian society in the American southwest, and the computed simulation, given information about climate patterns and so on, does roughly mimic what the archaeological record suggests really happened to the Anasazi.

The interesting thing is that the simulations, including this one, are really not much more sophisticated than Conway's famous "life" AI experiments -- they take a couple of crude populations and set up trivial rules, and then run with them until a pattern emerges. In spite of how crude these simulations are, the parallels to the observed world can be striking, suggesting that such simulations can be used to understand evolution, historical trends, racism, genocide, economics, etc.