glenn,
> You wrote:
> >>
> Medical students are often asked to explain the effects of mutations
> in examinations. All medical literature that I have seen refers to
> deleterious mutations, or at best, mutations that are marginally
> beneficial in rare circumstances and deleterious in others. Hence, is
> not your argument that "many combinations are possible" somewhat
> fallacious?<<
>
> No and here is why. I have a computer program which runs on a PC which
> models a being with an 8 unit genome. This 8 unit genome makes a unique
> screen shape. I can set the program to search for a genome that creates a
> particular shape (e.g. horizontal line, vertical line, diagonal line, disk,
> x or what ever.)
>
> The genome is mutated randomly. What is interesting is that whatever shape I
> ask for I can get ultimately after about 2000 mutations. After that, it
> becomes very difficult for any mutation to improve upon this solution. all
> mutations are "deleterious" to the preferred shape of the screen animal.
>
> But what is even more interesting is that everytime I run the program I get a
> different genome which gives the same phenotype on the screen!
> If I want a top left to bottom right diagonal line, thousands of genomes
> produce this figure! And they match very well. But once you find a
> soluiton, it is difficult to find a better solution and the longer you have
> searched for a solution, the less likely it is that you will find an
> improvement in your random search.
>
> Thus in living beings, which have been randomly mutated for countless
> generations, as long as the environmental pressures do not change, the
> best solutions have been found..
As a former computer programmer I am very well aware that it takes
intelligence to write a program. Without knowing more about your
program, I cannot comment further.
> You wrote:
> >>If the number of active alleles of which you are aware reaches a
> maximum of 59 in MHC, this multiplication by 59 of the chance of
> the random occurrence of these amino acid strings (nearly said
> nucleotide :-) ) is surely a minor, trivial and insignificant increase.<<
>
> You should have said nucleotides.
Why? The MHC complexes are proteins.
> The alleles are made of DNA but one can
But the proteins they code for are made of a.a. which simplify
without invalidating the argument
> equivalently look at the same problem in the proteins they code for. The
> problem is NOT the random generation of 59 different alleles. The problem is
> how, by the random mutation of the 10 alleles after the flood, can you give
> rise to the 59 observed alleles? This is quite a different problem than what
> you are suggesting. I am not suggesting the discovery of one of these
> alleles by random mutation but the generation of ALL these 59 alleles by
> random selection by modification of the existing allels. The pathway to the
> new allele must involve random mutation of the existing allele. The only way
> out of this is to assume that the flood was not anthropologically universal
> or that the flood occurred a long time ago.
Yeh, but that was not the point being made at the time.
> You wrote:
> >>Now is 59 in 10^130 any more likely? Okay, it's 59 times more likely,
> but it is still very unlikely. Change the a.a. string length to 102
> and the chance, 59 in n^102, drops to below 1 in n^100.<<
>
> Your calculation makes the mistake that all antievolutionary arguments make.
> You are assuming that ONE and ONLY one sequence will perform a given
> function, i.e. that you are looking for a needle in a haystack.
Read above which you just tried to refute. In the case of MHC
complexes, the number of known sequences that will perform the job is
59, which you claimed was the maximum number of functional alleles
known at any locus. I am talking a.a., even if there were six codons
for each a.a., the number is still astronomical.
> Let me illustrate this fallacy. A woman is born with 40,000 eggs and
> that is all she will ever have. A man produces 300 million sperm in a single
> ejaculation. To create you, a particular sperm must meet with a particular
> egg. Thus from your parents you represent 1/40000 x 1/ 300,000,000= 8.3 x 10
> ^-14. This is a single generation. But in order to produce your parents,
> two similarly chanced events must have occurred. Starting with your
> grandparents generation you represent a chance of occurrence of 5.7 x
> 10^-40. Going back 300 generations along one line only gives a probability
> that you will exist at 1 out of 10^-4200. But there are several million
> ancestors over this distance. thus assuming 10 million different ancestors,
> you have a chance of existence at 10^-29400. I don't think you exist!
> The problem is that while YOUR odds of existence (represented by your
> DNA) are quite small, the probability that SOMEONE'S odds of existence
> (represented by some DNA string) are quite high. If you want to calculate
> how likely is it that YOU will exist, it is highly unlikely. but if you want
> to calculate how likely it is that someone would exist it is much better.
> The same thing can be said of DNA, RNA and amino acid strings. Some would
> exist and there are millions of sequences of DNA which produce amino acides
> which perform the same function. Pig insulin is different than human insulin
> and yet it works (performs the function) in diabetics. the same with goat,
> cow, horse etc. The question you should ask is how many sequences can
> perform a given function? Billions of different DNA sequences perform the
> function of creating a human. That is why all 5 billion people on the planet
> have different, unique DNA sequences.
We are back to the above argument of the number of functioning
alleles. There are approx. 3.3 * 10^9 nucleotides in the human
genome.
i.e. 4^(3.3 * 10^9) possible human genomes.
2^(6.6 * 10^9)
2^10^(6.6 * 10^8)
> 10^3^(6.6 * 10^8) 2^10 approx. but > 10^3
10^(19.8 * 10^8)
10^(1.98 * 10^9)
Now what is the chance of any of the human evolving by chance,
according to the figures (5*10^9) you gave?
i.e. 5*10^9 / 10^(1.98 * 10^9)
5 * 10^( 9 - 1.98 * 10^9 )
Compare 9 to 1.98 * 10^9. Not much is it?
I have taken account of the known alleles.
> glenn
Graeme Cumming,
Medical Student,
University of Newcastle,
New South Wales,
Australia.
Email address: mi95gc@medicine.newcastle.edu.au