More non-random evolution

quote:

---

Originally posted by peter borger:
Dear percy,

​

Radioactivity is assumed to act randomly in the genome with respect to mutations. This is also expected for UV and oxidative stress. Apparently it is not random, as demonstrated by these 'intruiging' findings. What's up? Theory in trouble?
Best wishes,
Peter

---

I think you're misunderstanding the use of the term "random".If I roll two dice, and add them, I get 7 more often than 12. This doesn't mean the dice aren't random; it means that there is a structure within which these random events are being pooled.Imagine that you have an object which is weaker in some places and stronger in others. Now, hit it at a random point, and it may or may not break... Keep hitting it at random points. It is more likely that it will break at a weak place than a strong place, even though the places you hit it are random.This doesn't imply planning, design, or volition; it just recognizes that some structures break more easily than others.

quote:

---

Originally posted by peter borger:
Dear Seebs,

​

You say:
"This doesn't imply planning, design, or volition; it just recognizes that some structures break more easily than others."

​

I say:
"Exactly, and that may contribute to the alignment of mutations in DNA. Such mutations may look like common descent."

---

I don't think so at all. Imagine that any of a dozen mutations can occur at a given point. It would be *VERY* suspicious to claim that two creatures (A and B) each have the same sequence for the first 11 of them, and B and C have only 2 in common, but A and B don't have a common ancestor, they just *HAPPENED* to have such an unlikely sequence happen.If I deal two bridge hands, and someone ends up with the same set of cards both times, it's pretty hard to claim that the two starting sets of cards weren't similar.I think I see where you're going, and if there were only one suspiciously similar sequence of genes in one related species, it'd be a good enough argument to cast it into doubt. When instead we have hundreds upon hundreds of sequences across hundreds of species... That's different.It should be easy enough to test. Give one set of people access to a nicely detailed summary of the fossil record for a given chunk of the tree. Ask them to guess at relationships.Now give another people a set of gene maps, and ask them to guess at the relationships.If they come up with similar trees, we've got pretty good evidence that that's the best model.