In <Pine.SUN.3.91.950717090838.27936A-100000 at chuma>
karl at CHUMA.CAS.USF.EDU ("Stephen A. Karl") writes:
>>> As far as "junk DNA" is concerned, I think it is always good to
>> There are two types of non-coding DNA of unknown origin or purpose:
>> 1) Garbage DNA and
> 2) Junk DNA
I have been involved in evolutionary computation for the last couple of
years. As in biology, variable length EC genomes accumulate Junk Code.
Rencently, my colleagues and I looked at this phenomenon in EC and found
some strong indications that the "Junk" code in EC genomes plays an
important role in artificial evolution.
In our model, an example of junk code would be IF 2 < 1 then . . .
Our working hypothesis is that the "Junk" code plays a structural role
in allowing the genome to direct the attention of the crossover
operator to locations where high fitness blocks of code are not
disrupted. So far, ou
r results are consistent with this hypothesis. (The work is available
as a TR out of Univ Dortmund and will be availible in "Advances in
Genetic Programming II" due out sometime this year.)
I am quite interested in a biologist's perspective on this subject. Is
there any evidence that the junk code in DNA plays any kind of a
structural role in crossover, mutation? Is there any evidence, however
t the junk code plays any role whatsover?