A while back I remember wondering whether DNA might keep a record of the
edits and mutations much in the same way that programmers use RCS to control
versions of complex software. Sometimes a mutation will have immediately
obvious benefits or defects, but often the advantages may take many
generations to come out.

Suppose for example a certain gene makes an animal less attractive to a
potential mate, but confers a strong survival advantage under rare times of
stress such as an ice age. When the first ice age for 50,000 years hits the
species, it is possible that the gene will have become so rare that the
animal may become extinct. If, however, the animal has the ability under
times of stress to undo some of the mutations in the recent past - then it
has the option to revert to tried and tested states which may have survived
for much longer than the present one. This is not guaranteed to work, but it
is surely a lot better than making changes at random.

I was wondering whether the records of the edits might be the explanation of
the volumes of junk DNA that most species cart around with themselves. I
have since come across a better explanation.

Warren D Smith in DIMACS Series "DNA computers in vitro and in vivo" in
Discrete Mathematics and Theoretical Computer Science Vol 27 (1996) proposes
that complete Turing machines can be made out of RNA. The program and the
data is an RNA loop, and the machine that executes the program is another
piece of RNA that binds onto a site of the first loop, 'reads' the next RNA
sequence, 'looks up' the appropriate sequence in an internal 'matrix',
replaces the old segment with a new one, and steps an appropriate number of
sites along the loop, and starts again. He argues that this method is
similar to RNA editing in trypanosomes. He therefore argues that things such
as transposons, spliceosomes, and stuff like that are evidence of working
Turing machines, or at least the evolutionary relics of Turing machines.

He then invokes a computer science explanation for junk. Briefly it goes
like this if I understand it right. If the RNA Turing machine exists, then
it is easy to accidentally generate 'bugs' in the program that delete large
sections, or insert large strings of junk. More complex life forms could not
evolve unless there were some error trapping mechanisms (as yet unknown)
that arrest these fatal or burdensome processes before they insert or delete
too much. The junk he argues is the places where this has already happened.
There may be processes that remove this junk, but there may also be an
advantage to leaving it in...

Suppose the DNA is like a hard disc. If you have a few important files
scattered about the disc but most of it is empty, then the occasional
process that strikes a random sector of the disc will probably not be fatal.
If you have all your data packed together, then it is more likely that the
process will wipe several important programs before it can be stopped. The
junk DNA does not have to contain any error correcting stuff in order to be
helpful.

Me, I'm not a DNA expert. I expect I have got bits wrong. Read the original
article if you want the proper facts. However Warren Smiththe the nice point
that we know a lot about computers and little in detail about how DNA is
organized. Rather than using our knowledge of DNA to build better parallel
computers, maybe the flow of knowledge ought to be from computing science to
biochemistry.

Cheers.
Richard Kirk

PS:
General 'Stonewall' Jackson had this morbid fear that one arm was shorter
than another. He always used to ride with one arm in the air to make the
blood drain into the 'shorter' arm so it would catch up. He did that most of
his life, and it made no difference. Definitely one test-tube short of a
rack, that bloke

PPS:
I remember Sir Hugo Dolls posed the question "Why are our arms the same
length?". Arm growth is a chemical reaction that occurs over about 15 years
at about 37c, and then stops at the same point on either side to within
about 1% in most cases. There is no obvious feedback mechanism. Legs might
respond to the pressure difference if they were different lengths, but in
fact they don't - if you have one short leg, then it tends to stay that way
without corrective surgery. We know cells can probably count the number of
times they have divided, but this suggests something more fancy.





%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%