According to creationist and biologist Robert W. Carter, the reason that evolutionists need junk DNA is this: the alleged generation of genetic information is such a slow process that for the generation of our genetic material to fit into a feasible time line, most of our genetic material must be non-functional. In other words, evolutionists are attempting to simplify the scope of their problem by reducing the amount of genetic information they have to explain.
In Carter's own words:
Based on the work of J.B.S. Haldane and others, who showed that natural selection cannot possibly select for millions of new mutations over the course of human evolution, Kimura developed the idea of "neutral evolution". If "Haldane's Dilemma" were correct, then the majority of DNA must be non-functional. It should be free to mutate over time without needing to be shaped by natural selection. In this way, natural selection could act on the important bits and neutral evolution could act randomly on the rest. Since natural selection will not act on neutral traits, which do not affect survival or reproduction, neutral evolution can proceed through random drift without any inherent "cost of selection". The term "junk DNA" originated with Ohno, who based his idea squarely on the idea of neutral evolution. To Ohno and other scientists of his time, the vast spaces (introns) between protein-coding genes were (exons) just useless DNA whose only function was to separate genes along a chromosome. Junk DNA is a necessary mathematical extrapolation. It was invented to solve a theoretical evolutionary dilemma. Without it, evolution runs into insurmountable mathematical difficulties. [1]
Then he comes to the heart of the matter:
Junk DNA is not just a label that was tacked on to some DNA that seemed to have no function, but it is something that is required by evolutionary theory. Mathematically, there is too much variation, too much DNA to mutate, and too few generations in which to get it all done. This was the essence of Haldane's work. Without junk DNA, evolutionary theory cannot currently explain how everything works mathematically. Think about it; in the evolutionary model there have only been 3-6 million years since humans and chimps diverged. With average human generation times of 20-30 years, this gives them only 100,000 to 300,000 generations to fix the millions of mutations that separate humans and chimps. This includes at least 35 million single letter differences, over 90 million base pairs of non-shared DNA, nearly 700 extra genes in humans (about 6% not shared with chimpanzees), and tens of thousands of chromosomal rearrangements. Also, the chimp genome is about 13% larger than that of humans, but mostly due to the heterochromatin that caps the chromosome telomeres. All this has to happen in a very short amount of evolutionary time. They don't have enough time, even after discounting the functionality of over 95% of the genome--but their position becomes grave if junk DNA turns out to be functional. Every new function found for junk DNA makes the evolutionists' case that much more difficult.
So, as I understand it: the more functionality we find in our "junk DNA," the more difficult it becomes for evolutionists to explain how that genetic information could be generated inside the time lines that they laid out for us. How dreadfully inconvenient.
[1] Robert W. Carter. The slow, painful death of junk dna. Journal of Creation, 23(3):12-13, 2009.
Testing the commenting authentication system. Unfortunately it was necessary to require authentication due to a recent rise in spam attacks.