Researchers claim only ~8% of human DNA is functional

[Pythagorean]

Something I've always wondered about. It's neat to see a quantitative answer, finally.

"To reach the new figure, Dr Lunter and his colleagues took advantage of the ability of evolution to discern which activities matter and which do not.

They identified how much of our genome has avoided accumulating changes over 100 million years of mammalian evolution – a clear indication that this DNA matters, it has some important function that needs to be retained."

http://www.sci-news.com/genetics/science-only-8-2-human-dna-functional-02083.html

Not sure how robust their condition for 'functional' is though.

 

[gravenewworld]

"Functional" may be a bit of a misnomer, but we already know that the vast majority of DNA is not used to encoding proteins. The rest of the DNA contains information for things like miRNAs which are extremely important for gene regulation, but again, these regions only serve as a regulatory mechanism and not to greatly expand the amount of proteins possible.

We also now know that humans have less protein enconding genes than some strains of rice, yet who'd argue that rice is more complex than a human based on the sheer size of an encoding genome? The obvious argument people will try is that genes can be spliced in multiple ways to make different proteins. Well, the entire proteome, that is the grand total of all proteins , which of course reflects all of the entire output of the genome, is only about ~100K-200k proteins last time I checked, which can still be argued as still far too small to contain the complexity that defines a human. What people have ignored for a long time , however, have been the posttranslational modifications (PTMs) that get added to proteins after they're made. There are over 400 different known types of these modifications. Glycosylation of proteins is the largest class of PTMs. If you sum the total number of glycan structures that get added to proteins, you get what has been dubbed the 'glycome', which is akin to the genome for DNA and then proteome for proteins. The glycome is now known to be orders of magnitude more complex than the entire genome and proteome and in fact has been described as one of the most complex entities in all of nature. Carbohydrate structures on proteins can endow them with entirely new functions and can even supercede the importance of the protein itself for function. Unlike DNA and proteins , however, you can not control PTMs in a template like manner because there is no code that exists for controlling glycosylation like there is for DNA. The 20th century was dominated by DNA and genetics, but I'll put all of my money on the fact that the next revolution in biology and most of the effort in the next two centuries will be spent on deconvoluting and trying to figure out how to tame the realm of PTMs. PTMs are where the real molecular diversity occurs that produces the millions or even billions of distinct molecular species at any given moment in time that truly define life. Life is certainly sweet, and it is why there has been an explosion in carbohydrate biology research and why glycobiology has been named as one of the top 10 fields of science that will transform medicine in the 21st century by MIT. The trillions of possible combinations through PTMs is how such a small protein coding genome and small proteome can be transformed into a grand set of molecular complexity that is needed to create a human.

 

[Pythagorean]

The researchers are aware. The article mentions that of that 8%, only 1% was for coding, the other 7% was for regulatory functions.

 

[Ygggdrasil]

What fraction of the human genome is functional has been a fairly contentious issue for such a seemingly simple question. Part of the controversy comes from the fact that researchers from different subfields of biology have different ideas of what it means for a piece of DNA to be functional (see Doolittle et al. and Kellis et al. for further discussion). One of the first estimates came after the sequencing of the mouse genome in 2002 and, in comparing the mouse genome to the previously sequenced human genome, the authors concluded that it looked like ~ 5% of the human genome was under evolutionary constraint. Subsequently, this estimate has been continually updated as more genomes have been published and newer analytical methods have been developed, with most giving figures in the range of 5-10%. The PLoS Genetics paper that the OP's news article references, gives a number consistent with these previous estimates.

Interestingly, although only ~10% of the human genome seems to be conserved evolutionarily, about 80% of the genome seems to have some biochemical activity, such as binding to transcription factors or being transcribed. Some of this excess can be attributed to false positives (i.e. regions that are transcribed in low abundance that don't really have any function or regions where transcription factors bind non-specifically with no functional consequence) or the fact that some of the experiments were performed in cancer cell lines where biological functions are greatly perturbed. However, the methods to detect evolutionary conservation are far from perfect and they can have trouble detecting homology between species. Indeed, studies have demonstrated that these methods can overlook bona fide regulatory sequences. Furthermore, regions of the genome from diverse species can have http://gbe.oxfordjournals.org/content/5/3/532.abstract[/URL]. Thus, it would not surprise me if the number is actually larger than the 8% figure quoted above.

 

[Torbjorn_L]

Funny you should mention the mouse in a thread about genome functionality. Just for kicks, here is a lineage-specific genetic difference:

"What's mighty about the mouse? For starters, its massive Y chromosome"

"The mouse Y’s ancestral gene loss was even more profound than that found on the primate Ys. The mouse Y retains only 9 of its 639 ancestral genes, while the human Y held onto 19 of the more than 600 genes it once shared with its ancestral autosomal partner. However, the mouse more than made up for its early decay through rampant gene acquisition and amplification. Today, the mighty mouse Y carries 700 genes on a chromosome that dwarfs the human Y with its 78 genes. Strikingly, the region occupied by the acquired and amplified genes comprises nearly 97% of the mouse Y chromosome as a whole. Moreover, the Y represents a full 3% of the entire mouse genome."



[ http://www.sciencedaily.com/releases/2014/10/141030132955.htm ; my bold]

TL;DR: You don't want to mess with male mice.

 

[mes314159]

The comparative genomics studies described in the various replies do of course have another caveat, which is that they identify genomic regions required in at least most of the species being compared. If there are regions of our genome required in us but not in mouse (novel transcriptional regulatory regions for example), those will be missed by comparing our genome to the mouse genome. So the numbers in the 5-10% range are bound to be underestimates, though hard to say by how much. A different approach is to look at variation among living humans; presumably any site which varies in human populations is not absolutely essential for life (at least not in one copy, as we are diploid that makes interpretation harder). Or we can look at regions different between human and chimpanzee genomes; there too it's not easy to distinguish functional differences from random drift without doing experiments.