Hi Nick

Some interesting problems arise when we try to determine what's junk and what's not. Superficially similar species can have radically different genome sizes in the microbial world. Even amongst vertebrates there are weird anomalies - for example, the Queensland Lungfish, Neoceratodus, has a gigantic genome compared to us (and most mammals) but doesn't seem to do much with it. Some amphibians have similarly massive genomes. And when experiments have "purged" genomes of 'junk' the organisms have survived without obvious difficulties.

So just how much is non-gene, but still useful? In humans I've read that 5% is the non-junk region - 1.2% is genes, and 3.8% seems to be the regulatory regions the article discusses. Thus 95% appears to have no use at all - no surprise as much of it is "Alu" repeats, transposons, retroviral DNA, pseudogenes and similar 'clutter'.

But don't underestimate the importance of 'junk' - for example, transposons can both cause genetic disease through scrambling/deactivating genes, and can cause (potentially) beneficial mutations by carrying duplicated bits of genes with them when they 'jump'. Retroviruses are 'inactive' most of the time, but some duplicate in huge numbers as active viruses during the pregnancies of placental mammals. And 'pseudogenes' can be reactivated almost as easily as they can be deactivated.

So I'm a bit sceptical that it all has a 'function' or a 'purpose' in genomes, but it can still be important. What's really opening eyes in genomic medicine is just how many genes can have 'functioning' duplicates in any one person's genome, and how much they can differ. There's a lot going on that cheaper and cheaper gene-chip and sequencing technology will reveal in the next few decades.