Most eukaryotes posses a certain amount of junk DNA in their cell nuclei. What is (are) the origin(s) of this junk DNA, And is it realy junk (superfluous)?
"Junk DNA" is more aptly named noncoding DNA. This is defined as any DNA region that does not encode for a gene or more precisely is not within an open reading frame. In the human genome over 98% consists of noncoding DNA. However the more we learn about molecular biology the more we understand the biological function and importance of noncoding DNA. Examples for important functions are:
- Regulatory regions that control the expression of a gene
- Regions coding for regulatory RNA
- Regions where epigenetic regulation takes place
However, there are also regions which likely do not have beneficial biological function, which may rightfully be called junk:
- Transposons are genetic regions that can copy themselves (either by an enzymatically active RNA or by encoding for the protein transposase). They are believed to have evolved as "selfish genes" and several known defense mechanisms against rogue transposons exist (siRNA, RNAi). Transposons and the defense mechanisms have now become powerful tools in molecular biology research.
- Endogenous retrovirus sequences which are remainders of retroviruses which have inserted themselves into the germ line and become inactive through mutation.
However, even these "junk" regions are believed to have important evolutionary functions such as protection from mutation through retroviruses: Because there are large DNA regions where the precise order and function is not important, a retrovirus that inserts itself at random positions of the genome is less likely to cause permanent damage.
Briefly, we know of many mechanisms by which genomes can get larger. Tetrapods had at least two complete genome doublings in their history; transposons expand; retroviruses insert; partial duplications lead to pseudogenes. And these expansion mechanisms can be fast -- full genome duplications double size in a single generation.
But we know of very few mechanisms by which genomes can get smaller, and most of those are very slow, and very few are targeted.
From an mechanistic viewpoint, it's very difficult to imagine a targeted way to remove useless but harmless DNA quickly, and with 100% accuracy. If accuracy is not 100%, then the pathway would be more harmful than the DNA it seeks to remove.
The key is that if extra DNA is either harmless, or nearly harmless, there's no reason to eliminate it, and there are reasons (errors in removal) to not try to remove it.
So the short and simple answer is that genomes can accumulate useless DNA much more readily than they can get rid of it. It's just common sense, which matches 30 years of experimentation.