I thank Dr Poole for his thoughtful comments1 on my recent review2. In making these comments, he does not disagree with my thesis in general, but instead takes issue with some aspects of the supporting argument.

Dr Poole disagrees with the statement that "the amount of noncoding DNA rises as a function of complexity", especially the use of the word 'function'. This is partly a semantic issue, as the usage of this word is ambiguous, but it was chosen deliberately to be so, as I am suggesting in a general way that noncoding DNA (and the noncoding RNAs encoded therein) is important for developmental complexity. In any case, there is no doubt that, when ploidy is corrected for by such calculations, the relative amount of noncoding DNA broadly increases in line with increasing complexity, at least in those organisms whose genomes have been sequenced. This is only a crude measure and was not intended to be perfect, nor is the overall trend negated by exceptions, including unusual prokaryotes or eukaryotes with inordinately high or low amounts of repetitive sequences. The bacteria Mycoplasma leprae and Rickettsia prowazekii do have relatively large amounts of 'noncoding' sequences, but this is mostly a consequence of the degeneration of protein coding genes during reductive evolution to a more specialized niche (obligate intracellular symbionts)34. To select these examples as evidence that the link between noncoding DNA and complexity is false seems pedantic, and I would suggest that it is more a matter of being honoured in the breach.

Similarly, the fact that, for example, amphibian species might have huge amounts of repetitive sequences of unknown significance and functionality does not detract from the observation that more complex organisms contain, in general, larger relative amounts of noncoding sequences. Indeed Dr Poole appears to argue against himself, in that he posits that simpler eukaryotes that are (apparently) undergoing reductive evolution are losing noncoding sequences (introns), whereas those that are complex are not, but are gaining such sequences. The latter might or might not be functional, but the trend holds.

I agree with Dr Poole that many transposons and repetitive sequences might have acquired function, and I do make this point briefly in the last paragraph of the review2. In fact, the evidence is increasing that transposons are important contributors to the molecular genetics and dynamics of gene expression in complex organisms (for example, REFS 5,6), with the latest intriguing observation7 being that post–transcriptional A–I editing (which is most active in the brain) is vastly increased in humans and overwhelmingly occurs in Alu elements which are unique to the primates, whose neural capacity is most advanced. Like all immigrants, transposons might not be initially welcomed, but they become part of the community, change its dynamics and evolve along with it.

I am not suggesting that all noncoding DNA is functional. Clearly we have a lot to learn, but my general point is that the assumption that non–protein–coding sequences that have accumulated in the genomes of complex organisms are mostly evolutionary junk, and have remained that way, should be reassessed. At least 97% of the transcriptional output of the human genome is noncoding RNA and most of the human genome is transcribed, although less than 1.5% encodes protein8. I argue that the proportion of regulatory information increases with system complexity, and that in the case of complex organisms both logic and the available evidence indicates that most of this information is contained in non–protein–coding sequences of the genome and is transacted by noncoding RNAs2. Dr Poole's comments do not detract from this argument.