Are Reg deadlines too short for competent analysis these days?
This could have been an interesting and informative article, but some key errors crept in early on:
1) "Noncoding DNA" does NOT mean "not *associated with* encoding protein strings"; rather it means "not encoding protein strings". As an analogy, only the characters in this article that make up the article's words would be the "coding" characters, while the title, formatting, HTML tags, ads and so forth would be "noncoding" characters, even though many of them are in fact "associated with" the article's text.
The results of this major study have not changed our estimates of the proportion of the genome that is non-coding, as the coding part remains a small fraction of the total. Rather, the study has attempted to assign some function to much of the non-coding part that previously had no known function assigned. Again, pushing our genome-as-Reg-article analogy, the study has identified some characters as parts of the title, others as formatting characters, others as indicating where ads should be placed, others as ad content, others as the first characters of every line, and more. Clearly, the initial functions in this list are potentially quite interesting to would-be genome/Reg-article decoders. The later functions I listed, such as those involved with serving ads, are indeed "functions", but unlikely to be significant for the function of the article. Finally, there are likely to be chunks of characters without any function for the article -- perhaps a bit of someone's debugging core dump that was accidentally pasted in, or old, now irrelevant, comments. The last is what everyone would agree on as true "junk" DNA; many would also argue that the genomic analogs of ads and their functional elements also belong in the evolutionary "junk" bin (in DNA, these include sequences such as integrated retroviruses, transposons, and so forth). A controversial aspect of the recent study is that it also includes many physical features of the DNA as potential functional regions; although in a limited number of cases, these have been associated with some clear function, it's unclear whether the association holds in general. In our analogy, this would correspond to sites such as the first characters of every line. These account for much of the difference between the 20% and 80% numbers.
2) Nobody (other than the article's author) suggested that "functional" "...means that even being replicated counts as a 'function'". Since essentially 100% of the DNA in the cell gets replicated when the cell divides, that would unhelpfully define 100% of the DNA as "functional" just because it exists! At least one knowledgeable commenter did tweet that the 80% number "...includes definitions of 'activity' barely more interesting than 'replicated'..." and it's that definition of "interesting" that causes all the controversy.
3) Scientists should not decide what results to publish based on whether Intelligent Designists (or even so-called climate denialists) will find succor in their results. What is, is; if someone finds that to be compatible only with creationism, that's the person's own issue, not that of society. Along those lines, I'd caution readers that the mere fact of even 100% of the genome having some function would not be evidence one way or the other in that specific controversy.