A week ago, a huge, painstakingly orchestrated PR campaign was timed to coincide with multiple publications of a long-term study by the ENCODE consortium in top-ranking journals. The ENCODE project (EP) is essentially the next stage after the Human Genome Project (HGP). The HGP sequenced all our DNA (actually a mixture of individual genomes); the EP is an attempt to define what all our DNA does by several circumstantial-evidence gathering and analysis techniques.

The EP results purportedly revolutionize our understanding of the genome by proving that DNA hitherto labeled junk is in fact functional and this knowledge will enable us to maintain individual wellbeing but also miraculously cure intractable diseases like cancer and diabetes.

Unlike the arsenic bacteria fiasco, the EP experiments were done carefully and thoroughly. The information unearthed and collated with this research is very useful, if only a foundation; as with the HGP, this cataloguing quest also contributed to development of techniques. What is way off are the claims, both proximal and distal.

A similar kind of theory of everything hype surrounded the HGP but in the case of the EP the hype has been ratcheted several fold, partly due to the increased capacity for rapid, saturating online dissemination. And science journalists who should know better (in Science, BBC, NY Times, The Guardian, Discover Magazine) made things worse by conflating junk, non-protein-coding and regulatory DNA.

Biologists particularly those of us involved in dissecting RNA regulation have known since the eighties that much of junk DNA has functions (to paraphrase Sydney Brenner, junk is not garbage). The EP results dont alter the current view of the genome, they just provide a basis for further investigation; their definition of functional is biochemically active two very different beasts; the functions (let alone any disease cures) will require exhaustive independent authentication of the EP batch results.

Additionally, the findings were embargoed for years to enable the PR blitz at minimum unseemly when public funds are involved. On the larger canvas, EP signals the increased siphoning of ever-scarcer funds into mega-projects that preempt imaginative, risky work. Last but not least, the PR phrasing choices put wind in the sails of creationists and intelligent design (ID) adherents, by implying that everything in the genome has a purpose under heaven.

What did the study actually do? The EP consortium labs systematically catalogued such things as DNAase I hypersensitive and methylated sites, transcription factor (TF) binding sites and transcribed regions in many cell types. Unmethylated nuclease-sensitive DNA is in the open configuration aka euchromatin, a state in which DNA can discharge its various roles. The TF sites mean little by themselves: to give you a sense of their predictive power, any synthetically made DNA stretch will contain several such sites. Whether they have a function depends on a whole slew of prerequisites. Ditto the transcripts, of which more anon.

Lets tackle junk DNA first, a term I find as ugly and misleading as the word slush for responses to open submission calls. Semantic baggage aside, the label junk was traditionally given to DNA segments with no apparent function. Back in the depths of time (well, circa 1970), all DNA that did not code for proteins or proximal regulatory elements (promoters and terminators) was tossed on the junk pile.

However, in the eighties the definition of functional DNA started shifting rapidly, though I suspect it will never reach the 80% used by the EP PR juggernaut. To show you how the definition has drifted, expanded, and had its meaning muddied as a term of art that is useful for everyone besides the workaday splicers et al who are abreast of trendy interpretations that may elude the laity, lets meander down the genome buffet table.

Protein-coding segments in the genome (called exons, which are interrupted by non-protein-coding segments called introns) account for about 2% of the total. That percentage increases a bit if non-protein-coding but clearly functional RNAs are factored in (structural RNAs: the U family, r- and tRNAs; regulatory miRNAs and their cousins).

Go here to read the rest:
Junk DNA, Junky PR