More Recent Comments

Thursday, November 23, 2006

Bill Dembski Needs Help, Again

Bill Dembski asks, ...
I suspect that the “junk DNA” hypothesis was originally made on explicitly Darwinian grounds. Can someone provide chapter and verse? Clearly, in the absence of the Darwinian interpretation, the default assumption would have been that repetitive nucleotide sequences must have some unknown function.
Fortunately, there are some smart people who post comments on Uncommon Descent. They have told Bill that the concept of junk DNA is explicitly non-Darwinian. It was proposed by scientists who didn't feel the need to explain everything as an adaptation.

I don't know how many times we've explained to Bill that not all evolutionary biologists are "Darwinists." I know I first told him four years ago but I'm sure there were others before me. He seems to be a very slow learner.

One of these years he'll realize that there's more to evolution than just natural selection.

9 comments :

Dr. Andras J. Pellionisz said...

Questions (including the origin of the misnomer "junk") about the Zeitgeist of 21st Century ("Junk DNA") should be welcome.

Why not focus directly on the strictly scientific question of revealing its function?

While there are plenty of science issues at
http://www.junkdna.com
the most recent breakthrough (23rd of November) on diversity of human genome is a particulary pertinent "debate issue"; e.g. how the 98.7% of human Genome is responsible for the enormous human diversity - when 99.9% or our Genes are identical (and only the "beyond Genes" regions might determine our diversity).

pellionisz_at*no*spam_junkdna.com

Larry Moran said...

dr. andras j. pellionisz asks,

Why not focus directly on the strictly scientific question of revealing its function?

Why do you assume it has a function? All the evidence is against it and a proper understanding of evolution doesn't require that it have a function.

Anonymous said...

In genetic programing simulation runs with variable length genomes, junk "genes" will inevitably accumulate, often taking over 90% of the genome. It's an evolved defense against the destructive effects of the crossover operator. I've read that this may also be one reason behind junk dna (and introns) in biology. If true, it would be an adaptive "Darwinian" explanation - one which does not derive from the phenotype of the individual, but rather the stability of its offspring.

Larry Moran said...

jeffw said,

It's an evolved defense against the destructive effects of the crossover operator.

What does that mean? Homologous recombination isn't destructive (unless it makes a rare mistake).

I've read that this may also be one reason behind junk dna (and introns) in biology. If true, it would be an adaptive "Darwinian" explanation - one which does not derive from the phenotype of the individual, but rather the stability of its offspring.

This doesn't work as an explanation. I think it's based on faulty logic but you'll have to explain that "logic" in order for me to make a better judgement.

The yeast Saccharomyces cerevisiae has a high rate of homologous recombination but almost no junk DNA. Is that consistent with your explanation?

Anonymous said...

What does that mean? Homologous recombination isn't destructive (unless it makes a rare mistake).

In genetic programming models, the crossover operator (even though it increases variation) has the adverse effect of breaking up useful functional "groups" of software "genes", that work together. The accumulation of junk genes in the genome increases the probability that crossover will occur in groups of those junk genes and decreases the probability that working functional groups will be broken up. Thus genomes with junk dna are more fit (for the offspring). In GP, junk gene sequenes are called "introns", and GP crossover is usually not "homologous" to start out with. One book that discusses all this is Genetic Programming, an Introduction, by Banzhaf et al

Yeah, I don't know if this would be a significant problem in biology for homologous recombination (they're not exactly homologous) but the wiki seems to think so (third bullet under hypothesis and origin):
http://en.wikipedia.org/wiki/Junk_DNA.
Perhaps there was time when crossover was less homologous.

The yeast Saccharomyces cerevisiae has a high rate of homologous recombination but almost no junk DNA. Is that consistent with your explanation?

No it isn't. But I'm not suggesting that recombination is a sole cause. just a contributing cause, based on what I've seen in GP. If you could show that asexual organisms had huge amounts of junk dna, that would work against me as well.

Larry Moran said...

jeffw says,
No it isn't. But I'm not suggesting that recombination is a sole cause. just a contributing cause, based on what I've seen in GP.

It's very dangerous to draw conclusions about real biology from genetic programming. The algorithms in genetic programming are not good models of biological evolution.

The Wikipedia article lists several speculations about possible functions of junk DNA. (If any of them are correct then the DNA isn't junk.)

The one you're referring to is ..

Junk DNA may act as a protective buffer against genetic damage and harmful mutations. For example, a high proportion of nonfunctional sequence makes it unlikely that a functional element will be destroyed in a chromosomal crossover event, possibly making a species more tolerant to this important mechanism of genetic recombination.

This isn't a serious contender. Extra DNA will not protect against mutations and it will not "protect" against recombination since recombination is not destructive.

Furthermore, there's no way such protection against future events could evolve by natural selection. Evolution doesn't see into the future.

Anonymous said...

It's very dangerous to draw conclusions about real biology from genetic programming.

Agreed.

it will not "protect" against recombination since recombination is not destructive.

As a non-biologist, I'm not sure what to make of this. I've seen conflicting descriptions of recombination. Some of them say it leaves genes alone, and some say it can happen within a gene (If I remember right, The Selfish Gene says this). I even remember reading that it can happen within introns, so that function protein groups are preserved.

Furthermore, there's no way such protection against future events could evolve by natural selection. Evolution doesn't see into the future.

It most certainly does in GP, when crossover is destructive and the genome is of variable length.

Anonymous said...

...One more point: for this model to be meaningful, it doesn't have to be "destructive" in the sense that it breaks up a gene (cistron), just that it breaks up a functional group of genes working together (for homologous crossover, I don't know what that would mean).

In any case, here's the Selfish Gene reference: pg 28, 30th anniversary edition:
"..some people use the word gene interchangeably with cistron. But crossing-over does not respect boundaries between cistrons. Splits may occur within cistrons as well as between them."

Anonymous said...

"It's very dangerous to draw conclusions about real biology from genetic programming. The algorithms in genetic programming are not good models of biological evolution."
I see an out of context quotemine being snagged up in a hurry.