What do you mean by “non-coding”? Non-coding DNA I take to be DNA that does not undergo any transcription. Non-coding RNA would not undergo translation.

Either way, both ncDNA and ncRNA may have function and if they do are not “junk”, and we already know of many examples of each having some other function. Just because that function sometimes does not depend on the sequence does not make it “useless baggage”.

Let me predict that the number will be far closer to 93% than to 6%. Let me further predict that it will be possible to reverse engineer, so to speak, “junk” DNA to find the original, designed function.

I know that deletions _can_ happen at all scales, the issue is if they _do_ happen, and do the resultant organisms survive to reproduce. At first, I thought that deletion size plotted against frequency would follow a Poisson distribution. But now I am not sure.

If you look at the numbers on that Wiki page for diseases caused by different deletion events, you can see that some deletions, at least, recur frequently – much more frequently than a single base copying error!

The difficulty is that these deletions cut across both coding and non-coding regions. Because of the pattern of interspersal of coding and non-coding regions on the chromosome, once deletions get above a certain size, they are almost sure to take out a gene. As a result, the cells of the organism aren’t making all the proteins they need. This is usually fatal.

The issue with prokaryotes is that they do not have “junk DNA” in the same way as eukaryotes. So a general theory explaining junk DNA has to explain the differences between eukaryotes and everythng else.

The Wiki page on “Junk DNA” links to this recent article on the size of ‘required’ non-coding regions. The basic idea of the article is that eukaryote architecture solves a connectivity problem by using non-coding DNA to create non-coding RNA, and it is possible to calculate how much ncDNA is needed for this purpose. The answer for a human size genome is about 6% of ncDNA. So the ultimate answer on how much ncDNA is uesful to the organism, and how much is excess baggage, will probably fall simewhere between this number (6%) and the 93% quoted by PaulN from the ENCODE preliminary results.

Why do you insist that we are only talking about “base by base” deletion? This is not what is meant by a deletion mutation.

http://en.wikipedia.org/wiki/Deletion_mutation

“Any number of nucleotides can be deleted, from a single base to an entire piece of chromosome.”

I see no reason why this is somehow restricted to prokaryotic cells either. There are single-celled, eukaryotic organisms as well, and I’m quite sure their mutation and replication rates are similar to prokaryotes. As for multicellular, complex eukaryotic organisms such as mammals, the difference is only one of scale.

As you have said,

“The answer would seem to be that some really is junk, and is therefore decaying away, while some other part is serving some function and being conserved.”

I don’t think that any of it is junk. If it is decaying away, it is only doing so because natural selection was not strong enough to weed out mutant individuals, allowing non-functional mutants to propagate slowly throughout the population. You wouldn’t call extinct species “junk” would you? Was the dodo “junk”? Were the dinosaurs “junk” species? Are all the animals on endangered species lists junk? I think they function just fine. If not, why try to save them?

“We can see how useful it would be to cull non-functional DNA from the genome, and reduce the energy budget of the cell. But if there is no source of variation that regularly deletes large swaths of non-coding DNA, it will never happen in nature.”

I never suggested it was a single mutation that did this. Multiple mutations deleting small swaths of useless DNA would add up. We are talking about DNA in general here, which has been in existence since the beginning of life on earth. There is plenty of time for multiple deletions to do the job. Perhaps not for recent organisms with slow mutation and reproductive rates, but the genetic material from whence they came would have already been established before then.

Appendix?

“Junk” and “useless”, when applied to something we ourselves did not have a hand in building, designing, creating, or otherwise bringing into being, are more likely to indicate a failure to understand on our part than a failure to design well on the builders part.

I agree, but base by base deletion is not going to eliminate 90% of the DNA in the genome. To accomplish the large efficiencies under discussion, you need some wholesale process. Just for a reference point, I looked at Allen Macneill’s list of sources of variation to see if it included such a method. It did not. At least in eukaryotes, it looks like the best we can do is turn things off, but leave the code in place.

Ok, going back over the OP.

Mr PaV got the attribution slightly wrong. The comment is by someone posting as “superhuman”. “Flag” is a link to report the comment as abuse or spam.

In any case, the comment is, I think, incorrect. Selection can only act on the variations that are present. If there is no mechanism to create that variation, it can’t be selected.

We can see how useful it would be to cull non-functional DNA from the genome, and reduce the energy budget of the cell. But if there is no source of variation that regularly deletes large swaths of non-coding DNA, it will never happen in nature.

This means that whether the protein itself is actually deleterious or not doesn’t matter. If the protein doesn’t do anything, the genome is likely to lose the gene for it in a short amount of time.

It’s very easy to imagine. Mutations can cut out entire genes quite often. If the organism doesn’t need that gene, then there is no selective pressure to weed out the mutated organism. In fact, the organism is spending less energy than the wild type, so it may even provide a selective advantage to lose genes that don’t do anything.

This has happened in the lab with microorganisms. Grow E. coli on a particular medium for long enough, and it will lose the genes it uses for other food sources. Why? It doesn’t need those genes because it’s getting all it needs from one medium.

Ah, do you mean “how long can either gene stay in some mutated, non-functional state”? I think that the answer depends on selection pressure, or if the products of the mutating gene are actually deleterious. For example, a duplication of opsin genes might be responsible for some of our color vision. If the duplicated gene started making poison, not opsin, then its population share will drop fast. I don’t know enough population genetics to give a better answer.