In article <3dtcql$a10 at kaiwan009.kaiwan.com>, bshapiro at kaiwan009.kaiwan.com
(Benjamin Shapiro) wrote:
> want to write a computer model in C to figure out the probability of the
> creation of a protein in the "beginning of earth"
There is a problem here, due to the fact that proteins probably were not
the first biomolecules. Current theories favor a prebiotic RNA world.
There is an even more severe problem here, due to the fact that
probabilities of the appearance of anything in the course of evolution
critically depend on the process of selection.
As an illustration: consider your target should be a sequence of 100 A s
(_A_lanine). What is the probability of this specific sequence arising
through random combination of our usual 20 aminoacids ? 1/(20**100) ! If I
remeber correctly, the current estimate of the number of particles in the
universe is negligible by comparison to the number of tries you would need:
about 10**29 (or 20**22.3).
Now lets model this as an "evolutionary" process:
Initialize with a random sequence of 100 amino acids (characters)
Loop, until all characters are A
randomly change any character * (MUTATION)
if this changes a character from anything to an A * (SELECTION)
then accept the change
discard the change
A few thousand iterations will get you there. Try this a couple of times,
maybe trying to get the sequence of real proteins from random sequences.
THEN vary the selective forces (i.e. allow the introduction of "wrong"
characters with a small probability). Note how the number of necessary
iterations goes up. You can graph number of iterations vs. selective force.
You can also work out the result analytically: eg. how many iterations
would one need to get the desired sequence with a probability of 99% under
the given constraints.
The point I'm trying to make is: it all depends only to a small degree on
the actual prebiotic conditions, but very much more on the nature of the
Boris <steipe at lmb.uni-muenchen.de>