This is a simulation of genetic drift over 700 generations. 200 bottles are filled with 64 flies of the species drosophila melanogaster. 48 of the 64 flies have the genotype br/br (homozygous brown), 16 are wild type Br/Br.
The simulation shows that without natural selection, the genetic variety shrinks. At the end of the simluation, all populations either are brown (br/br) or wild type (Br/Br). Heterozygous individuals have disappeared. Note that the ratio of br/br and Br/Br populations is equal to the initial ration in the first generation.
This simulation has been created using python, gdchart bindings, ImageMagick and mencoder
@LeopardFrogPilboxhat
So it makes sense that you heard of it going the other way, this is a special setup which differs from what happens in nature.
It is possible to re-run the simulation from a specific generation onwards. However, I'm not sure what to change. IIRC, a certain amount of flies is chosen randomly to interbreed. That means that the process is not deterministic, if you run the simulation twice, you don't get exactly the same numbers. HTH
papopapolo 2 years ago
@LeopardFrogPilboxhat
Yes, the alleles get eventually fixed. It can take some time but at the "end", there are no more heterozygous individuals left.
Note that this is a special setup where natural selection is inhibited/not present. If you do the experiment as described (really run it, wit actual bottles etc., not just the simulation), you get similar results. This has ben done to discard evolutionary models which do not account for natural selection.
papopapolo 2 years ago
Question: Do your drift simulations always tend to remove diversity?
Because I've heard of it going the other way
It would be interesting if you were to test some historical contingencies. Could you do that? could you run and log a simulation, then go back, change one thing and resume? "re-run that tape of life" as gould said
That would be a useful video.
Thanks for making this one though
LeopardFrogPilboxhat 2 years ago