Originally Posted by MalFet
Makes sense for evenly distributed alleles (.5^10 is mighty small, after all), but I've still got to think that this has the tendency to eliminate relatively less frequent traits (.9^10 isn't quite so tiny). If we did this with people, we'd run out of redheads pretty quick.
Well, the idea is obviously that genetic diversity isn't even ideal with yeast. You're taking 10 colonies of (virtually) genetically identical yeast cells. The aim is to keep the yeast the same, with minimal change. So not only is the disappearance of redheads not a problem, but you don't want them pop up, and when they do, you WANT them to be stamped out. The point is, that with good practice, the traits you want should *never* be relatively infrequent within your sample.
You can't really compare yeast with people like that, because there's an enormous, key difference - genetic diversity is completely undesirable when you're trying to maintain a pure yeast culture.