In their recent book A Scientist’s Guide to Talking with the Media (which I enthusiastically recommend, by the way), Hayes and Grossman describe the role that university press offices have in disseminating new findings by researchers at their institutions. I agree that this is an important job and that good press releases can have a very positive effect. The corollary, of course, would be that poorly crafted ones can sow confusion. I have been critical of science blogs and science news services in the past, but in some cases they are simply re-posting (albeit uncritically) the stories from press offices, which may be where the actual problem is.

Witness two frustrating examples from genome biology. The first was by the press office at Johns Hopkins [How Neutral Genetic Drift Shaped Our Genome], and was re-posted by some science blogs. The second is by the University of California, San Diego, and is re-posted at ScienceDaily and Scientific Blogging [One Man’s Junk May Be A Genomic Treasure].

Both stories are guilty of over-hyping the significance of the research (which perhaps is not surprising) and of including significant factual errors (which is not acceptable). Notably, the Johns Hopkins release mangles basic evolutionary theory, and now we have this from UC San Diego:

Scientists have only recently begun to speculate that what’s referred to as “junk” DNA — the 96 percent of the human genome that doesn’t encode for proteins and previously seemed to have no useful purpose — is present in the genome for an important reason. But it wasn’t clear what the reason was. Now, researchers at the University of California, San Diego (UCSD) School of Medicine have discovered one important function of so-called junk DNA.

The first line is patently false. In my area of study, I encounter far too many speculations regarding functions for non-coding DNA, and this is how the situation has been for decades. It also bears noting that the study in question studied one transposable element, SINE B2, which makes up around 2.4% of the mouse genome and appears to contribute to the regulation of a growth hormone gene. This is not very surprising; recall that McClintock first characterized transposable elements as “controlling elements”, and even the earliest and most vocal proponents of the “selfish DNA” hypothesis surmised that some TEs would have regulatory functions. SINE B2 itself has been implicated in regulation at least since 1984 (see also here from 2001). This is not in any way a critical comment about the work — it is sure to be an interesting study and I look forward to reading the article when it appears in Science this week. But this press release — which I suspect had little to do with the authors of the study — is vastly overstated to the point of twisting the history of the discipline. (They also suggest that protein-coding genes make up 4% of the human genome whereas the real total is less than 2%, but that’s a comparatively minor issue).

I am very interested in working with the media to provide accessible, interesting, and (not or) accurate information to the public. I also realize that writing about scientific research is difficult, and that there are many individuals out there who are very good at it. I don’t like to be critical all the time, but we really must clean up reports on “junk DNA”. I am open to any suggestions on how scientists can help to make this a reality.

More later, but just a quick post to give credit where it is due. Kudos to New Scientist for their excellent story “Why ‘junk DNA’ may be useful after all” in the July 14 issue. Aria Pearson gets the complexity of the situation right while making the piece interesting and accessible.

I have some quick questions for the scientists, journalists, and neither-nors who read Genomicron. Should scientists nit-pick inaccuracies in news reports? Does it do any good to keep journalism on track, or is it a waste of time (or maybe even irritating)? What else can we do to improve the accuracy of science reporting, in particular involving cooperation with rather than criticism of science writers? Since I don’t have the answers yet, I will just go ahead with a snarky post in the meantime.

Here is what triggered this query. I was reading this story, How Neutral Genetic Drift Shaped Our Genome, at Scientific Blogging (based on a Johns Hopkins press office release), when I came upon this gem:

When they expanded their study across the whole human genome, they found more than 1200 such pieces of mitochondrial DNA of various lengths [nuclear pseudogenes of mitochondrial origin, or numts] embedded into chromosomes. While chimps have a comparable number, mice and rats only have around 600 numts. Since they increase in frequency as species advance, it suggested there was some evolutionary purpose to keeping them around.

Strikingly, however, none of these numts contained the blueprint (an actual gene) to make a protein that does anything, nor did they seem to control the function of any nearby genes. “At best, it seems numts are a neutral part of our genome,” says Katsanis. “If anything, they may be mildly negative since long repeat sequences can be unstable or get inserted inside genes and disrupt them.”

Ok, the nit-picker in me really wants to point out that evolution is not a process under which “species advance”, that there is no such thing as an “evolutionary purpose”, that genes are not blueprints, that this study deals with a tiny fraction of the genome, that the idea that “they accumulate steadily” makes no sense in terms of either function or neutrality given what they just said about rodents versus primates, and that of course these pseudogenes — of mitochondrial descent, no less — do not encode a functional protein by definition. There is absolutely nothing surprising in the observation that they evolve neutrally in the nuclear genome.

To Larry Moran’s recent post Stop the Press!!! … Genes Have Regulatory Sequences!, we can add Pseudogenes do not encode functional proteins! Next: DNA includes both protein-coding genes and non-coding sequence!

(Update: Scientific Blogging redeems itself with a good post about the ENCODE hype here)
(Another update: Evolution Diary uncritically posts the same nonsensical press release here.)
(Another update: The paper itself is actually far more interesting than this poorly crafted press release would make it appear. See the open access pdf here.)

I used to subscribe to (and quite enjoy) The Scientist when it was a free publication, but have mostly stopped reading it now. Two stories in the latest issue were pointed out to me by readers of Genomicron, so I thought I should provide some brief comments.

The first relates to an editorial by Richard Gallagher entitled “Junk Worth Keeping“, which asks “Is it time to retire provocative descriptors such as ‘junk DNA’?”. The answer, says Gallagher, is “no”. Why? Because “junk DNA” is useful for… framing. He cites Nisbet, although as I have come to learn, framing, sensu Mooney and Nisbet, applies to conveying science in a particular way to help win an election. I’m not sure what “junk DNA” has to do with defeating the Republicans as per Mooney’s formulation of framing, and in any case, Greg Laden has already provided a very useful discussion of the fact that “junk DNA” is a frame that anti-evolutionists exploit because it allows them to obfuscate genetic knowledge to their own ends. They do, of course, get both the history and the science of the term totally wrong, but that’s not the point when it comes to spin. In other words, the term “junk DNA” is a horrible way to communicate the complexity of non-coding DNA to an audience with no background in evolution or genome biology. Granted, it does get them to talk about the issue, but I believe there must be better ways to do this than to set up the discussion to be oversimplified and confused.
[Hat tip: junkdna.com]

The second comes from a story by Melissa Lee Phillips reporting “Surprises in the Sea Anemone Genome“. Here are some statements of interest:

1. “…the sea anemone, one of the oldest living animal species on Earth…”
2. “The study also found that these similarities were absent from fruit fly and nematode genomes, contradicting the widely held belief that organisms become more complex through evolution.”
3. “It’s surprising to find such a ‘high level of genomic complexity in a supposedly primitive animal such as the sea anemone,’ Koonin told The Scientist.”

I will probably do an entire post on the common fallacy of describing one extant lineage as older than another (common descent dictates that all contemporary lineages are of exactly the same age, although some have undergone more branching and/or morphological change than others). But statement #1 is especially inaccurate because the age of a species is not the same thing as the age of a lineage (which is what the author meant to describe). This species (Nematostella vectensis) could, for all we know, be very young, and certainly there is no basis in evidence for calling it one of the oldest species on Earth. (The original paper in Science is at least clearer on this, describing the Cnidaria as “the oldest eumetazoan phylum”, though this is difficult to substantiate given that no unambiguous fossils exist for eumetazoa prior to the Cambrian, at which time early representatives of modern phyla were already present; they cite two papers that can, at best, only suggest that Cnidarians [and perhaps molluscs] were present among the Ediacaran biota). And, of course, “the” sea anemone is a misnomer as this name applies to the entire order Actinaria, within which there are several dozen families.

Statement #2 may refer to a widely held belief among the public (and perhaps among some genome sequencers), but the expectation that complexity always increases in evolution is a fallacy that was abandoned long ago by evolutionary biologists. (It also bears noting that organisms do not evolve, populations do). A huge amount of the diversity on this planet is composed of parasites, and their evolution often involves simplification. In this case, it is not morphological but genome evolution that is under discussion, and there is really no reason to expect an increase in complexity here either. Gene loss is a well known phenomenon, and in fact one could make a case that streamlining of the genome or having one gene do multiple things is related to increased morphological complexity whereas gene number is not. Take, for example, the sea urchin, whose immune system seams to involve a large number of genes, as compared to a vertebrate, whose immune system generates nearly endless variation by recombining a comparatively limited number of genes. (Incidentally, the anemone genome appears to contain 18,000 genes).

Statement #3 is not necessarily inaccurate, but I caution readers that it must be interpreted in a certain way. “Primitive” does not mean “less complex” or “less advanced”, it means “more like the last common ancestor of the groups being compared”. In this sense, a sea anemone is probably more like the last common ancestor of all animals than is, say, a fly. But this is also a modern species whose overall lineage has been around for exactly the same amount of time as that of the fly. The anemone lineage may have undergone fewer morphological changes (though still probably a lot) than the lineage that led to insects, but the anemone itself is not the common ancestor and in fact may bear only a modest resemblance to the ancestor. This is exactly the same confusion people face when thinking about humans and chimpanzees. Chimps were not the ancestor of humans, and may bear only a modest resemblance to the common ancestor of the two lineages — the two split millions of years ago, and both lineages have undergone considerable change since, with many species arising and disappearing in the meantime.

I like to see these interesting topics covered in The Scientist, but I must admit that these two pieces in particular do not inspire me to re-subscribe to the magazine.

While I’m at it, let me register a small complaint about the story by Elizabeth Pennisi in Science [Sea Anemone Provides a New View of Animal Evolution]. I have become mostly resigned to the fact that science writers will insist on describing genome sequencing as “decoding” a genome, and this story is no exception. What bothers me more is that Pennisi is so sloppy in characterizing evolution as a “progressive” process. Thus, she argues that “genome sequencers have just jumped down to a lower branch on the tree of life“, and that “until now, researchers have relied heavily on the sequenced genomes of the fruit fly, nematode, and that of a few other invertebrates to understand genome evolution leading up to the vertebrates“. (Update: Apparently the authors the paper cautioned Pennisi about this, but to no avail!).

Evidently, we have a lot of work to do in clearing up the basic details of how evolution occurs.