More Recent Comments

Wednesday, June 20, 2018

Press release from the Francis Crick Institute misrepresents junk DNA

Press releases have become a serious problem. I'm frequently upset whenever I read a press release covering a field I'm familiar with. They rarely do a good job of explaining what's actually in the paper and putting it into the proper context. The people who write press releases are more concerned with sensationalizing the work than they are with teaching the general public about how science works. They often do this with the blessing and participation of the scientists who did the work.

Let me illustrate the problem using a recent examples from the Francis Crick Institute in London, UK [Non-coding DNA changes the genitals you're born with]. The press release covers a recent Science paper from the Lovell-Badge lab ....
Gonen, N., Futtner, C.R., Wood, S., Garcia-Moreno, S.A., Salamone, I.M., Samson, S.C., Sekido, R., Poulat, F., Maatouk, D.M., and Lovell-Badge, R. (2018) Sex reversal following deletion of a single distal enhancer of Sox9. Science. [doi: 10.1126/science.aas9408]
These workers discovered and characterized a regulatory region upstream of the mouse Sox9 gene. The Sox9 gene controls the development of testis and deletion of the regulatory region reduces the level of Sox9 gene expression leading to XY individuals that are phenotypically female.

We have known about regulatory DNA for more than 50 years so the paper doesn't make any contribution to our general understanding of transcriptional regulation. In fact, it fits right in with decades of work on enhancers, promoters, and transcription factors.

You wouldn't know that from reading the press release. Even the title of the article (Non-coding DNA changes the genitals you're born with) suggests that there's something unusual about noncoding DNA that has a function. This point is highlighted in the press release ...
Only 2% of human DNA contains the 'code' to produce proteins, key building blocks of life. The remaining 98% is 'non-coding' and was once thought to be unnecessary 'junk' DNA, but there is increasing evidence that it can play important roles.
This is 2018. Isn't it about time that science writers stopped spreading this fake news? There was never a time when knowledgeable scientists thought that all noncoding DNA was junk. Never.

Furthermore, we've had a pretty good understanding of regulatory DNA since the early 1980s. Think about what that means. It means that "increasing evidence" is a misleading way of saying that the basic facts have been known and understood for more than thirty years. Forty years ago you might have gotten away with saying that there's "increasing evidence" that noncoding DNA has a function, but not today.

Is this just sloppy science written by an employee who really doesn't understand the history of gene expression and genome composition? No, it isn't just ignorance on the part of the press office because they have the support of the lead author on the paper; a postdoc named Nitzan Gonen. She is quoted in the press release ...
Dr Nitzan Gonen, first author of the paper and postdoc at the Crick, says: "Typically, lots of enhancer regions work together to boost gene expression, with no one enhancer having a massive effect. We identified four enhancers in our study but were really surprised to find that a single enhancer by itself was capable of controlling something as significant as sex."

"Our study also highlights the important role of what some still refer to as 'junk' DNA, which makes up 98% of our genome. If a single enhancer can have this impact on sex determination, other non-coding regions might have similarly drastic effects. For decades, researchers have looked for genes that cause disorders of sex development but we haven't been able to find the genetic cause for over half of them. Our latest study suggests that many answers could lie in the non-coding regions, which we will now investigate further."
Here's a better way of explaining the significance of this paper.
The opening sentence of the paper says, "The regulation of genes with important roles in embryonic development can be complex, involving multiple, often redundant enhancers, repressors, and insulators."

These regulatory elements are usually found near the genes they regulate and they represent an important part of the genome. This study identifies a regulatory element that controls the Sox9 gene in mice. Defects in regulatory elements are known to cause genetic disorders and it has long been suspected that disorders of sex development are also due to mutations in regulatory elements. This study identifies an important regulatory element that controls sex development and demonstrates that mutations in this element cause sex development disorders.
There's no mention of "noncoding DNA" in the paper and no mention of junk DNA. That's because nobody is surprised to find regulatory elements that aren't in coding exons. Nobody who reads the paper is going to be surprised to learn that noncoding DNA has a function even though they understand that 90% of our genome is junk. Why can't the press release make this clear to the general reader? Why can't the authors make sure the press release accurately represents the published report?


25 comments :

PZ Myers said...

Because a lot of authors don't understand the concepts?

Unknown said...

I talk briefly with the author, she know the facts, but she under estimate the public understanding.

qetzal said...

“Our study highlights the important role of what some still refer to as ‘junk’ DNA.”

Sorry, that’s just false. No halfway knowledgeable biologist ever considered enhancers to be junk DNA.

Now, maybe she was misquoted. But if not, then either she doesn’t know the facts, or she knowingly made a false statement that over-exaggerates the significance of her work.

Larry Moran said...

If the public doesn’t understand the concept of regulation and junk DNA then it’s up to scientists and science writers to correct that misunderstanding, not pander to it. This press release just contributes to more misunderstanding.

strangetruther said...

A thought just occurred to me. Is it possible for a piece of DNA from which no RNA ever gets produced, to act as a "regulatory element"? Presumably it might if it acts as somewhere for the reading machinery to attach to?

John Harshman said...

Yes, that's what transcription factor binding sites are.

strangetruther said...

Thanks John, interesting. Yes, I've heard about them (taught about them too! The old cytochrome-C thing in OU SD226!) I've only just realised they don't produce any RNA, so if you're looking for non-junk by checking all the RNA that emerges from the nucleus, that won't find them. Presumably you have to search upstream of whatever they affect, to find them. I guess that might sometimes be tricky.

Anonymous said...

It is pretty embarrassing that some still continue to stick to the junk DNA nonsense. Doesn't the time pass for the retiree crowed?

Robert Byers said...

Hmmm. on the point about press releases in science or anything. Is it a intent to sensationalize the subject being covered? They would say they must stress the importance of the subject and why its covered relative to others.
they would deny they sensationalize . They get rewarded in the MEDIA for gaining audiences and maintaining them. YES they desire the exciting point!
Easily they don't reread their first impression.
it seems in this one of the researchers backs up the press release and so they would say DON'T blame us!
only a few people understand these things and the public never can weigh the issue. Creationists deal with this all the time. We deal with a public that must be shown they can understand and question concepts very easily.

Unknown said...

This Atlantic article, The Mysterious ‘Jumping Gene’ That Appears 500,000 Times in Human DNA, based on a paper in Nature, describes a stretch of DNA, called LINE1, which appears to be clearly both functional and junky. Sure, it does something, but it probably doesn't need half a million copies.

Mikkel Rumraket Rasmussen said...

Also known as promoters.

Mikkel Rumraket Rasmussen said...

Yeah it's the old divide by 500.000 problem. Technically one single LINE1 insertion could have a function, as a promoter for a downstream gene, or it might sit downstream from a promoter and be transcribed into a functional RNA, and the entire 499.999 rest of them could be nonfunctional. But the article makes it seem as it all of them have been found to be functional. Nothing is said about whether few or all of them are functional, so it makes it seem as if it is all of them.

Unknown said...

I think there are two concepts of junk DNA out there. One is DNA that lacks function and one is DNA of unknown function. Nobody ever thought enhancers were functionless but for many years nobody knew how to find them either because unlike coding sequence there was no easy way to identify them. Now there is. Using concepts and techniques developed by ENCODE I might add.

John Harshman said...

Yeah, there are two concepts, but one of them is wrong. Guess which.

Larry Moran said...

@Chris Campbell

You would be hard-pressed to find any experts who ever said that DNA of unknown function was junk.

Enhancers and other regulatory sequences were not only found but extensively analyzed in the 1970s and 1980s. My own lab did a lot a promoter bashing in the 1980s to identify regulatory elements. There were approximately 3000 other labs doing the same thing. :-)

The standard technique developed in the 1980s was to put a bit of DNA in front of a reporter gene to see if it had enhancer activity. Gonen et al. used a reported gene (hsp68) derived from a gene cloned in my lab back in 1986.

DNase I hypersensitivity was discovered by Hal Weintraub and Mark Groudine back in 1976. They were working in the lab next door to where I got my Ph.D. Techniques to identify sites of methylation and histone modification were worked out in the 1980s and 1990s. ENCODE researchers improved these techniques and applied them to the entire genome but they did NOT discover the concept and they did not invent most of the techniques.

Unknown said...

Larry, as often happens I think I expressed myself poorly. I assume (and if I'm wrong please correct me) that when you say no one ever said that DNA of unknown function was junk, you are saying that function can be inferred to exist even if the function itself is unknown if the DNA sequence in question has been evolutionarily conserved. What you're (and presumably the other experts are) referring to as junk DNA is DNA for which function can't be inferred because the primary sequence is poorly evolutionarily conserved. Have I got that right?

Larry Moran said...

@Chris Campbell

By the end of the 1960s it was obvious that many genomes were much larger than necessary to contain the expected number of genes. There was much speculation about the possible function of this excess DNA. It soon became apparent that much of it was repetitive DNA and that the sizes of genomes in very similar species could vary by a factor of ten.

When these facts were combined with Neutral Theory and genetic load arguments, it was reasonable to conclude that much of the “excess” DNA had no function. It was junk.

My point is that there was never a time when experts simply called all of the excess DNA junk just because they didn’t know what it’s function was. They had good reasons for believing that it didn’t have a function. The concept of junk DNA was never an argument from ignorance in spite of what it’s opponents would like you to believe.

Today we have lots of data that allow experts to infer that 90% of our genome is junk. They can identify function for about 5% of the genome. As you point out, there’s lots of conserved DNA of unknown function. Nobody thinks that fraction has to be junk. Furthermore, there’s probably lots of nonconserved DNA that will turn out to have a function. Nobody is saying that it all has to be useless junk. They are simply saying that there’s an upper limit to the amount of our genome that can be functional and that limit is less than 10%.

This is why papers that discover a function for some new bit of DNA are not surprising to junk DNA proponents. The do not refute the claim that 90% of our genome is junk. It’s extremely annoying when these papers are promoted as evidence that most of our genome is functional. It’s extremely annoying when otherwise intelligent scientists make such a silly argument. I assume they are doing it because they don’t understand the evidence for junk DNA. I assume they don’t believe that most of our genome is junk so they see their data as conformation of their (uninformed) bias.

anonymous said...

Regarding

...Furthermore, we've had a pretty good understanding of regulatory DNA since the early 1980s.

Is that perhaps somewhat too eukary-centric?

Jacob & Monod elucidated the Lac Operon in the 60’s

OK... polycistonic operons do not occur in Eukaryotes, but just the same...

John Harshman said...

And repressors.

anonymous said...

Convergent Evolution of Alu & Line could to some minds hint how parasitic DNA can be co-opted into functionality and that functionality could be “multifunctional”

Regulatory & Bulk are not mutually exclusive to this way of thinking

Just throwing that out there for the sake of argument

Brian said...

@Chris Campbell

"Nobody ever thought enhancers were functionless but for many years nobody knew how to find them either because unlike coding sequence there was no easy way to identify them. Now there is. Using concepts and techniques developed by ENCODE I might add."

Phylogenetic footprinting was pioneered by Webb Miller and colleagues at Penn State, and others at other institutions, in the early to mid 1990's precisely as a means to identify highly conserved non-coding sequences in gene loci that could possess regulatory function (sequenced manually from cosmid clones on P32-labeled gels back then, I might add). When tested as Larry mentioned in reporter assays, this strategy proved accurate in identifying transcriptional regulatory activities. The limitation to that point had been slow processor speeds that led to the multi-pairwise alignments taking days. An exponential trend in increased processor speed from that point on allowed this strategy to continue apace. The ENCODE project is the contemporary expansion of these pioneers' methods.

Unknown said...

I want to thank Dr. Moran for educating me about 'junk DNA'.
One conclusion I have come to that I feel completely certain about is this-

The notion that 'junk DNA' comes from an 'argument from ignorance' is an ignorant argument.

Unknown said...

@Brian
Phylogenetic footprinting won't allow you to pick out an enhancer that is active at a particular time and spacial location during development, especially when said enhancer is over 500kb upstream of the gene it regulates.

SRM said...

To be more exact, promoters (or promoter regions) are sites to which RNA polymerase binds to begin the process of transcription. Promoter regions may also include one or more transcription factor binding sites where some of these TFs may be activators and some may be repressors. There can also be other nucleotide regions within promoters that do not bind TFs or RNAP but that can still influence the rate of transcription initiation.

Brian said...

Chris Campbell: "Phylogenetic footprinting won't allow you to pick out an enhancer that is active at a particular time and spacial location during development, especially when said enhancer is over 500kb upstream of the gene it regulates."

Sure it can. If a specific non-coding sequence is constrained by an essential function (e.g. transcription factor affinity) at any developmental window or in any tissue, by any target gene in the genome, it can be identified by its multi-species conservation. This is precisely why the approach works (since all tissues have the same genome to use at all stages of development). Defining the function and target promoters of the conserved elements identified in silico with wet lab approaches is the fun part that comes next.