(All of these links appear to be behind subscription walls. It may be possible to get to the Science blurb using BugMeNot .)
Science Now talks about an idea emerging from comparing genomic sequences: basically there is an awful lot of conservation outside of coding DNA. This must be important stuff, but the function of this non-coding DNA is not yet clear.
The mouse (2.5 Gigabasepairs) and human (2.9 Gbp) genomes have a lot more DNA sequence than is used to directly specify protein (these are so-called coding DNA regions). In order to understand the function of the non-coding regions of the human genome, computational scientists look at which sequences are conserved between mouse and human (and other studies even compare human to fish, or fly). The result of these comparisions is that the total extent of non-coding sequences conserved between mouse and human is larger than that of coding sequences , and some "ultraconserved" sequences are actually better conserved than the protein-coding regions. .
The specific effort cited in Science actually shows that the abundance of these highly conserved non-coding DNA regions increases with the complexity of the organism pair being compared. When comparing yeast to roundworms, about 40% of the matches are non-coding; but when comparing fruit flies to vertebrates almost 2/3 of the conserved sequences are non-coding.
It's thought that these conserved non-coding regions are enhancer elements , meaning they contribute to the control of nearby protein coding regions. The idea would be that to make a more complex organism you'd need not just fancier building blocks (proteins) but also more precise control of where and when the proteins get put in. See also the references here in Pharyngula's reading list.