Skip ahead to the ENCODE stuff if you don’t care for the Tomkins posts.
For his blog post for this week (or last, depending on how you look at it) Jeffrey Tomkins wrote Deleting “Junk DNA” – Does it Matter? I’ll let him explain some of the background:
Does deleting “Junk DNA” in laboratory studies, such as in mice make a difference? Interestingly, a colleague of mine just recently pointed out a paper in which exactly this type of research was undertaken. In fact the study is a few years old, and was done in 2004. However, after a search of the literature, I have not been able to find anything similar.
Because this paper, however, shows up prominently on the web as some sort of proof for “Junk DNA”. I felt that a brief review of the facts that were actually discovered in the research are in order.
I don’t think I’d ever come across the paper – Megabase deletions of gene deserts result in viable mice (pdf) – before, actually. But it does look like it could be useful in future. Let’s see if Tomkins can persuade me against it, shall we?
First, a few brief comments on “Junk DNA” are in order. Less than 5% of the human and other mammalian genomes, like mouse, contain DNA that directly codes for protein. These protein coding segments contain the areas of chromosomes that are traditionally recognized as genes. Because scientists originally did not understand what the non-coding parts of the genome did, they labeled it as junk, hence the term “Junk DNA”.
Once again, we have the ‘all noncoding DNA is junk’ idea, which is nonsensical. Nobody thinks that, and it’s fairly certain that nobody thought it either. As an interesting aside, the paper under discussion doesn’t actually use the term “junk DNA,” instead claiming to “support the existence of potentially ‘disposable DNA‘ in the genomes of mammals.” (Emphasis added.)
However, over several decades of research have shown that the non-coding parts of the genome contain a wide diversity of DNA sequence that for the most part, plays a key roles in the development, growth and physiology of an organism. In fact, Dr. Jonathan Wells recently published an outstanding book covering the whole range of issues associated with this field of genetics called “The Myth of Junk DNA”. I highly recommend this book to anyone interested in this issue.
Tomkins really likes The Myth of Junk DNA, and More Than a Monkey uses it as a citation in, I think, more than half of its chapters. ‘Recently,’ in the case of that book, is “May 2011.” It’s certainly true that many functions for non-coding DNA have been found in specific cases, but there’s not nearly enough of that to dent the portion of the genome (the vast majority) that has no known function and isn’t really expected to have one.
The introduction concludes with:
Highlighting the importance of non-coding DNA discovery in the news for 2012, I also recently published a blog post describing how a certain class of retrolements [sic] are key in the initial stages of mouse embryo development (link).
That link goes to a reposting of the second DpSU covered in this post.
With all that over with, he begins:
The 2004 paper in question is titled “Megabase deletions of gene deserts result in viable mice”, which is unfortunate, because it is somewhat miss-leading. One segment of deleted mouse DNA was 1.5 million bases and the other was 845,000 bases in length. The first impression from the title of the paper makes it sound like they were making multiple megabase-sized deletions all over the genome. Nevertheless, these are two large chunks of DNA. The mouse genome in total is similar to the size of the human genome at about 3 billion bases.
It’s not all that misleading: plural is plural, after all, and ‘megabase’ is the correct order of magnitude here. That doesn’t make it the perfect ‘proof of junk DNA,’ but it’s good evidence.
These segments of mouse DNA were chosen because they contained areas of the genome called “gene deserts” – meaning they were completely devoid of not only protein-coding gene sequences, but also contained very few transposable elements and other active regulatory DNA sequence. Interestingly, these two deleted regions did, however, contain over 1,200 “ultra-conserved elements” highly similar to their counterparts in human. See my previous post on ultra-conserved elements and their anti-evolutionary properties here.
Ultra-conserved elements you say? Bullshit.
The abstract to the paper states:
Together, the two deleted segments harbour 1,243 non-coding sequences conserved between humans and rodents (more than 100 base pairs, 70% identity).
The definition used in the paper cited in Tomkins’ previous post was “stretches of at least 200 base pairs of DNA that match identically with corresponding regions in the mouse and rat genomes,” although more lenient (i.e. only 100+ base pairs) definitions are also used. 70% doesn’t cut it.
What’s more the paper proper is explicit on this matter:
Together, the two selected regions contain 1,243 human-mouse conserved non-coding elements (>100bp, 70% identity), also similar to genome averages, while no ultra-conserved elements or sequences conserved to fish are present.
Yes, you read that correctly: there were no ultra-conserved elements. Not even the one – and certainly not 12,000. While we do have less-conserved elements, they’re not quite the same thing and aren’t the problem that their more identical brethren would indeed be if they were present here.
Ultra-conserved elements are profound evolutionary enigmas – profound Darwinian DNA contradictions. They are highly similar (conserved) chunks of DNA across mammals (e.g. humans, rats, mice), but presently appear to show very little function (some minor enhancer activity). Hence they should be highly variable between major types of mammals, since very little hypothetical selection is acting upon them (neutral DNA).
You can read my earlier post for the various problems with that paragraph. Moving along:
Nevertheless, the researchers chose these segments to delete using biotechnology DNA tools because of their lack of supposed functionality. They eventually created strains of mice homozygous for the deletions, but did not perform long-term evaluations of multiple generations. To test the effects of deleting the DNA, the researchers observed both embryological development and mice for 25 weeks after birth. Parameters tested included visual appearance, growth rate, organ pathology, and blood plasma profiles. Apparently all of the visually evaluated parameters and blood profiles were normal.
It is a little unfortunate that a longer-term examination was not done, but this criticism is fairly weak. If these 2 million bases are as functional as the ought to be, should all DNA be functional, then it would have been noticed in the development stage when an appreciable proportion of gene expression (along with the “key roles in the development” that non-coding DNA is supposed to have). Instead we have all results normal, so far at least.
The researchers also evaluated expression changes (RNA profiles) for 9 genes flanking the deletions for 12 different organs. While a number of expression levels showed possible effects, only two were statistically significant. In fact, they were markedly different for two genes, one in the brain and the other in heart tissue. Since these are two of the most critical organs in the body, the regions that were deleted obviously played some major role of importance and were not just useless chunks of DNA. Also keep in mind that only 12 organs were tested and the researchers did not state the age of the mice tested. Presumably they were adults. Gene expression does vary markedly with the developmental stage of an organ.
It’s possible that Tomkins wasn’t paying that much attention when reading this paper. He complains that “the researchers did not state the age of the mice tested.” However the paper says:
We then carried out visual and pathological examinations on multiple organs from the various groups of mice at 6 months of age, including brain, thymus, heart, lungs, liver, spleen, stomach, intestine, kidney, urinary bladder, uterus, ovaries and testicles. No morphological abnormalities, evidence of abnormal growth or tissue degeneration were observed in delMm3 / delMm3 or delMm19 / delMm19 mice. Organ mass was similar in both groups of deletion mice and their wild-type littermates.
Finding no differences in any of the whole organism phenotypes tested, we subsequently explored the impact of the deletions at a molecular level. The expression levels of multiple genes flanking the boundaries of the two deletions were determined in 12-week old mice.
In other words, yes they do say how old the mice were (6 months in the case of the organs, i.e. the end of the experiment). I’ll add that thirteen organs are listed there – Tomkins said they did only twelve – and the paper implies that more were tested.
Now, those genes with the altered expression. Continuing on from where the above quote leaves off:
We assayed four genes bracketing the Mm19 deletion and five genes bracketing the Mm3 deletion by real-time quantitative PCR, in a panel of 12 tissues representing the overall expression patterns of each assayed gene. The tissue specificity of expression for all the genes tested was similar in homozygous deletion mice compared to their wild-type littermates (Fig.3). Out of the 108 quantitative expression assays (12 tissues for 9 genes), only 2 revealed detectable alterations in levels of expression. The expression of Prkacb was reduced in the heart of delMm3 / delMm3 mice and Rpp30 was reduced in intestine of delMm19 / delMm19 mice, compared to wild-type littermates (Fig. 3).
I may have missed where they explicitly call the changes ‘marked,’ although figure 3 does suggest that that description may be justified (at least assuming that the error is not unfavourable).
Going back to the above-quoted paragraph, Tomkins said:
Since these are two of the most critical organs in the body, the regions that were deleted obviously played some major role of importance and were not just useless chunks of DNA.
His logic is flawed here: it doesn’t really matter where the genes were expressed. Indeed, if a major change in gene expression in a critical organ doesn’t cause any noticable difference beyond that then what is observed may in fact go against what he says in that sentense.
All that being said, the question of how these changes are being caused still needs to be answered. There may be no genes in this region but there’s still the potential for some kind of regulatory DNA. The obvious candidates are those not-quite ultra-conserved elements already mentioned:
While non-coding conservation in the Mm19 desert was shared only amongst mammals, the Mm3 desert contained human-mouse conserved elements also shared with chicken (75 elements) and frog (5 elements). From the Mm19 desert, we picked five human-mouse conserved elements representing the most conserved sequences between humans and mice (>180bp, 90% identity) for the in vivo assay. The ten elements chosen from desert Mm3 (>400bp, 90% identity) included all five sequences that are conserved across humans, rodents, chicken and frog, and five that are conserved among humans, rodents and chicken only (Fig. 4).
We cloned each element in a reporter vector 15, injected it into fertilized mouse oocytes and assayed for the presence of beta-galactosidase activity in the resulting embryos at 14.5 days post-coitum . Eight to sixteen independent transgenic mice that were generated for each of the 15 elements were examined . Of the elements tested, only one, located within the desert Mm3, reproducibly drove beta-galactosidase expression in a set of tissues that include mammary glands and abdominal muscles (Fig. 4). It is noteworthy that this element is conserved deep in the vertebrate lineage, including human, mouse, chicken and frog.
So the area isn’t a complete desert. But it still seems to mostly be one, and I’m sure the mechanism behind the observed changes could easily be packed into this single element – the 2 million bases are unneeded.
The effects of the DNA deletions were also not done under natural conditions where the mice would be living in the wild. Also, the mice were not stressed in any way – often genes only kick in under certain stresses and stimuli. Additionally, did the brain expression changes affect the mice mentally – how would you test that?
The researchers do give the following disclaimer: “In assessing the impact of these deletions on the engineered mice, it is important to acknowledge that our ability to phenotype an organism will always miss some features no matter how detailed. It is possible, even likely, that the animals carrying the megabase-long genomic deletions do harbour abnormalities undetected in our assays, which might impact their fitness, in some other time scale or setting than the ones assayed in this study.”
You can test behavioural differences, which is a good enough proxy for ‘mental’ changes. It should be very clear by this point that Tomkins has little go on – only potential changes, not things that have actually been observed. Avenues for further study, yes, but by no means substansive attacks.
The bottom line is that gene expression was altered in these mice in very important organs – the brain and heart. Perhaps these areas of the genome where not essential to life, but they were part of God’s created design. Although we have five fingers on each hand, all fingers are not essential to existence, but they are key to an optimal existence. The same is true for nearly all regions of the genome.
It its fairly obvious, however, when you don’t have all your fingers, unlike what was observed here. One possible ‘use’ for junk DNA is simply to sheild more necessary regions of DNA from mutations. That too would be “not essential to life, but … key to an optimal existence.” And yet it would still be ‘junk,’ by any reasonable definition. Would Tomkins have a problem with that? The above paragraph suggests ‘no,’ but I suspect he would nonetheless. It would not reflect well in his Designer, and that’s what is key here.
He concludes, at last:
Far from showing the uselessness of certain non-coding regions of the genome, this study actually shows how even deleting seemingly gene desert regions has an effect on gene expression. Also, the fact that these regions contain the evolutionary enigmatic pieces of DNA called ultra-conserved elements further indicates that something important is contained therein. Contrary to the hype, this research is just more evidence for intelligent and optimized design in living systems.
You’ll notice the return to the (fictional) ultra-conserved elements. Tomkins case is weak, and he is relying on them to reinforce it. If the elements were present then yes, they would be significantly problematic – though not for the reasons that Tomkins gives, leading me to suspect that he does actually know their true nature and significance here – but they are not and the merely ‘conserved’ elements are poor substitutes.
In conclusion, this paper is no proof for junk DNA – this isn’t mathematics after all. But it remains solid evidence, and one I thank Tomkins for bringing to my attention.
This post was written some days ago, but delayed as it wasn’t urgent and I wanted to publish the Wednesday DpSU first. As it turned out, I never managed to get my brain working properly yesterday so that post was rather poorly written, but that’s neither here nor there. For the explanation of the implications of the delay in publication of this apparently ‘non-urgent’ post, I’ll have to ask you to Please Turn Over.