Posted: Jun 15, 2017 2:31 pm
by DanDare
Borrowing from Dan Dennett the set of all possible genomes is Vast (capital V = hugely humungous beyond imagining).

If you were to randomly pick a genome from that Vast set the chance that it would work in a cell with all the other bits of cellular machinery is Tiny (capital T = so small you couldn't find it this side of the plank unit).

Fortunately, biology doesn't assemble random genomes. It assembles copies of existing, functioning ones. Over and over and over. The chance of a functioning DNA strand being copied correctly is reasonably good. Mistakes occur. But machinery in the cell tend to eliminate very bad mistakes.

The mutations that get through can be randomly part of a large set of mistakes that can be made from a particular DNA strand, but the number of possible alternately arranged copies of a strand of dna is Tiny compared to the Vast potential set. All of the single error variations could be tried out by having different mutations in a billion different cells replicating once every hour over a ten to a hundred year stretch.

Of those variations a large number, more than half, probably less than 90%, are useless and wont work any more, eliminating them from the set of next steps, and halting the genome search down that branch. Of the remainder most of the errors make little difference to the new organism and so those variations will be available as the base for a next set of variants. Some small number may be changes that improve the creatures reproductive potential and will tend to become more common than the other variations.

The statistics are easy for life and evolution to deal with. Its just not that much of a mystery.