The Finch and Pea

This month there has been a bit of a dust-up over the question of how much of our genome is functional. ENCODE results say 80% – or do they? Is it 20%? Or more like 8%?

Did ENCODE scientists play fast and loose with the definition of function, or is genome function legitimately defined as those activities the consortium measured? Is functional DNA something that has an effect on phenotype? (Does that include damaging gain-of-function mutations?) Is functional DNA only that DNA present in your genome because of natural selection? (Then what about hitchhiker alleles?) Is a novel mutation existing in only a single individual functional if that mutation is ultimately destined to become fixed in the population by natural selection?

We have to face the fact that, like much else in biology, boundaries between categories are fluid. It makes no sense to try to cleanly divide the genome into functional and non-functional elements. Even what seems like an obvious boundary line, the boundary between protein-coding and non-coding DNA is blurry: many coding regions have cis-regulatory sites with a non-coding, functional role. To divide the genome into categories of coding- and non-coding function, or function and non-function, may satisfy our insatiable desire to classify for our own cognitive comfort, but from the perspective of the cell there is no such distinction. Continue reading “The non-functional concept of genome function”

I’m not sure how many of the people writing Science news features, press releases for ENCODE*, or completely uninformed and baseless rants on the idea of junk DNA are familiar with the work discussed in this review, none of which is refuted by the ENCODE results:

“The ecology of the genome — mobile DNA elements and their hosts”, John F. Y. Brookfield, Nature Reviews Genetics 6, 128-136 (February 2005):

One activity of evolutionary biologists involves looking at features of organisms and seeking to explain them in adaptive ways — demonstrating that the feature to be explained will confer on its bearer a higher inclusive FITNESS than an alternative would. However, as applied to phenotypic features, this approach is not always intellectually rigorous — only knowledge of the ways in which genes influence the phenotype can allow the identification of realistic alternatives to observed traits. This approach is more valid when applied to genomic components — an explanation of the presence of a DNA sequence consists of demonstrating that an organism with that sequence is fitter than one that lacks it or one in which the sequence is mutated. The methodology is straightforward — we make mutations and observe the reduction in fitness that is created. All parts of the genome could therefore potentially be seen in this same light — every sequence present is there because its removal or replacement would cause a reduction in the organism’s fitness. In discussing microorganisms, such a view might be tenable. However, the genomes of multicellular eukaryotes possess sequences, which could perhaps form the majority, that are not there for reasons related to their present use.

Why does a simplistic view of an entirely functional genome fail? In essence, it does so because some genomic components, notably interspersed repetitive DNA sequences, are indistinguishable from parasites…

This paper develops the ecosystem analogy of the genome. Later this week, I’ll develop the analogy of your genome as a post-apocalytpic wasteland.

*Sadly, a significant number of ENCODE scientists seem completely unaware of this literature as well.