Continuing a chapter-by-chapter review of Gauger, Axe and Luskin's Science and Human Origins,
a Discovery Institute publication that is intended to
challenge--amongst other things--the notion that humans share a common
ancestor with chimpanzees, and that we couldn't have had descended from a
literal Adam and Eve.
---
Yesterday, I changed the format of my review a little. I
predicted a few things about what I was expecting from Casey Luskin's effort to dismantle the argument for junk DNA - Chapter 4 - before I had the chance to read it. Let's see how I did. I'll also mention now that this chapter review is going to get technical at times.
Luskin's chapter is entitled "Francis Collins, junk DNA, and chromosomal fusion". Apart from trying to make the case against the existence of junk DNA, Luskin wants to make it clear that he is very disappointed in Collins, who he chooses to frame as "an evangelical Christian who embraces both Darwinian evolution and embryonic stem cell research". I can only imagine that the reader is expected to disapprove of the last point. The interplay between Collins' faith and science is an issue for Luskin as "his emphatic defense of ape/human common ancestry still has wide influence in the faith community." Collins might infect the faithful with science and lead them away from human exceptionalism. Luskin also notes that Collins and "atheist Darwinist Richard Dawkins" say very similar things when they speak about evolution. I think the implication there is pretty clear.
When I was thinking about what to expect from this chapter, I predicted that Luskin would conflate non-coding DNA and junk DNA, and that Luskin would exploit this erroneous conflation by pointing to known functions of non-coding DNA as evidence against junk DNA. Here, Luskin wastes hardly a word. In the midst of his anti-Francis-Collins tirade, Luskin points out "studies have found extensive evidence of function for non-coding DNA showing that it is not genetic “junk” after all." He then starts a section called
Non-coding DNA: Not really junk after all, evidently keen to drive home the point. In here he cautions the reader that "even a cursory review of the scientific literature shows it is wildly inappropriate to simply assume that repetitive DNA—or most others types of non-coding DNA—are useless genetic 'junk.'"
Bloody hell.
I'd like just once to see all these references where all these researchers are saying that if DNA does not code for a protein then it is junk. The whole issue leads me to wonder how much of the relevant literature Luskin has actually read. In any serious, introductory discussion of junk DNA I'd expect to see Ohno's 1972 paper introducing the concept of junk DNA referenced, and Ohno's argument discussed. Luskin does no such thing. He simply wants us to believe that 'Darwinists' everywhere irrationally assert that all non-coding DNA is junk DNA. Well, the only assertions are his, and the dozen creationists before him that have done the same. None of the quotes Luskin provides from Francis Collins come close to saying any such thing.
Indeed, knowledge of some of the functions of non-coding DNA predates the term 'junk DNA' in the literature. In other words, the argument for junk DNA
has always started from an understanding that non-coding DNA can be functional. It just happens to also be the case that a large swathe of the typical mammalian genome -- and ours is decidedly typical -- shows no evidence of function, and actually could not be evolutionarily conserved by natural selection, for reasons
I've explained elsewhere already. The strawman junk DNA proponent who simply assumes the existence of junk does not exist.
Luskin moves onto discussing transposable elements. These are highly repetitive sequences that collectively make up close to half of our genome, and so are at the heart of any discussion about junk DNA. Like non-coding DNA in general, Luskin lists some known functions for transposable elements and other repetitive elements like satellite DNA. In fact, he lists more than an entire page of bullet-pointed functions. Surely, if there's a whole page of bullet points, then transposable elements are functional. Right?
To answer that we need to consider what transposable elements are -- something Luskin clumsily omits from his chapter. Transposable elements get their name from their unsual property: they 'jump' around in the genome. Certain classes of them also make copies of themselves that reinsert at random in the genome. When either of these things happen, it is similar to other types of mutation -- some function might change, or they might happen to land somewhere non-functional and cause no harm. But, like other mutations, variations in these transposable elements between different people are often linked to disease -- in other words, sometimes transposable elements disrupt a genetic function and cause harm and disfunction.
Because we can examine the genomes of different people and see where new copies of these transposable elements have been made, we get a pretty clear picture of how good they are at self-replication in our genome, and so we also get a pretty clear picture that they've all come from this activity. The most prolific transposable element has more than a million copies in your genome! Even though sometimes a new copy might land somewhere and shift the function of a gene, or later mutate and become part of a new gene (and there are certainly recorded cases of this), the majority of these duplications have no effect. Mostly, they land between genes or inside introns where they change nothing of importance. If that were not the case, most of the duplications of transposable elements would either have to be removed by purifying selection, or would be causing disease.
So, you have large tracts of your DNA that have come about from the replication of genetic elements that can copy themselves. This would lead to an exponential increase in copies of these transposable elements if they continued to do so. However, because it is rare for their sequences to be performing functions in our genome, they also mutate freely and quickly. The typical transposon accumulates random mutations to the point where it no longer is able to self-replicate. We break them with mutation. Of the about 45% of your genome made up of these repeated copies of transposable elements, less than 0.1% are functional retrotransposons.
Just to be clear, let me repeat this: sometimes these duplications fortuitously land somewhere and are functional, conferring soem benefit. There are a few well-known examples -- I discussed a really awesome one at the end of a
two part series on junk DNA I've previously written here at
Still Monkeys. However, when we consider a) the origin of such sequences, b) that we would expect many copies of such dulpications to be retained in populations if they are neutral (i.e. they don't worsen fitness) and c) we are talking about half of the entire human genome, then the burden of evidence for function falls on the side of the 'no junk' brigade.
Luskin continues with an attempt to provide such evidence. Now is a good point to recall one of the other predictions I made: that Luskin would provide qualitative evidence of function in intergenic and other non-coding DNA, but will fail to provide a quantitative assessment for how this impacts our view of the genome. That is to say, he'd point to newly discovered function where none was known before and claim victory, yet fail to inform his readers that these interesting discoveries actually only account for tiny fractions of the genome. He does not tell his readers that because the fractions of explained DNA are so small, the proportion of purported junk DNA in mammalian genomes has not shifted since the very first assessment by Ohno, now 40 years ago.
An area that Luskin highlights is that of pervasive transcription. When a gene is used, it is transcribed into RNA. Those RNA transcripts might then be translated into proteins (this is what happens for coding sequences) or else they might be directly functional (e.g. regulating genes, forming ribosomes and so forth). Luskin is right to emphasize the diversity of roles played by RNA, as they have been somewhat underappreciated in the past.
Transcription is the first step when a stretch of DNA performs a cellular function. What Luskin latches onto is that several studies have suggested much of our genome gets transcribed into RNA, pointing to the possibility that a far greater proportion of the genome is functional than had been previously considered. Luskin highlights a number of commentaries (although virtually no primary research literature) on this side of the argument, doing his readers a serious misservice. This is, in fact, a highly contentious area and there are very good reasons to temper our excitement over pervasive transcription.
Many of the RNA transcripts that have been detected from areas of the genome that lack recognised functions are low frequency transcripts. It is widely recognised that because of the way transcription is initiated, there are many spurious transcripts -- "transcriptional noise" or junk RNA -- that are degraded by cellular processes such as nonsense-mediated decay. In yeast, Struhl (
2007) argues that 90% of transcripts are spurious. And while certain techniques have detected high levels of transcription in humans, other techniques that are
less error-prone have failed to do so (van Bakel,
2010,
2011) and have been the source of continuing controversy (Clark et al.,
2011). WIth this said, previously unidentified, functional RNA transcripts will continue to be identified. Hundreds of non-coding RNAs have been identified being expressed in specific tissue types, which suggests they may be functional (Mercer et al.
2008), although no function for them has yet been identified. But, let's remember to compare those hundreds with the tens of thousands of identified genes in the mammalian genome that collectively sum to only about 2 or 3% of the genome. In the vastness of the genome there is still little evidence of pervasive function, and there remain good reasons to expect much junk despite the evidence of transcription.
Luskin bothers with none of this massive grey area, none of the controversy. He instead moves onto pseudogenes. Pseudogenes are genes that no longer function in their former capacity because they have been degraded by mutation. Luskin highlights one example from humans used by Francis Collins as evidence of common descent, which Luskin simply calls a vitamin C pseudogene, and describes as being "supposedly functionless". Okay - is Luskin about to prove that this gene (
GULOP) is not actually functionless? No -- he moves onto a general discussion about other pseudogenes instead. A pity. GULOP is fairly compelling evidence of our common descent amongst the primates, as Collins has previously said. The haplorrhine primates (that means us and our closer relatives) differ from the other primates in that at one point, mutations fixed in GULO, the last component in the vitamin C synthesis pathway. Most likely, this is because the fruit-based diet of the common ancestor of the haplorrine primates was enriched with vitamin C, such that the loss of the pathway was not detrimental. The same thing has happened in some fruit-eating birds. We retain the legacy of this pathway in our genome, but none of its function. This non-functioning legacy is shared by the other closely related primates. The single mechanism that plausibly explains the sharing of a once-functioning, but identically broken gene is that we are related by descent. The only alternative is to insinuate that it is only "supposedly functionless" without a single bit of evidence, which is Luskin's approach.
Some pseudogenes get transcribed as RNA and sometimes act as regulators for genes. Such pseudogenes are not junk, and there are a couple known examples. We have about 20,000 pseudogenes, so again this is a numbers game. The majority are non-functional, and contribute to our total junk (although they total only about 1% of our genome).
I predicted that Luskin would talk about introns and alternative splicing. Introns break up our genes with non-coding sequences. All of our protein-coding sequences make up less than 2% of our genome, but the introns that break them up make up about a third of our genome. However, introns are not discussed at all by Luskin. This is a major omission.
The rest of the chapter is spent discussing the fusion of two ancestral chromosomes to form human chromosome 2. Luskin argues that this is not proof of common descent, and there may not have even been a fusion event. He argues that the telomeric DNA in human chromosome 2 is shorter than the telomeres found at the end of typical chromosomes. Again, this is designed to cast a small shadow of doubt on our common descent with other primates. No positive argument is offered for an alternative model.
This is what also struck me in Luskin's discussion of transposable elements. In a chapter of a book purportedly about intelligent design, I would have appreciated a model under which such selfish replicating elements in our genome could be understoood as a rational component of design. Perhaps this would tell us something about the nature of the mysterious designer. For example, a designer who would use such elements would appear to be a hands-off designer, because they would be willing to trade the suffering of dysfunction and disease caused by these elements with occasional adaptation over the long-term. An oddity about ID is that its proponents dislike anyone trying to draw inferences about the designer, yet I have never heard any complaints about archaeologists using the designed objects of ancient civilisations to do the same.
I haven't given any of the population genetics arguments for why there must be either junk DNA, or at least DNA that can freely accumulate change. These are a compelling, positive basis from which to understand junk DNA. The reason I haven't discussed them is because Luskin failed to address them, including the original one by the originator of the term 'junk DNA'.
I've explained elsewhere already for those who'd like to know more than what Luskin is willing to tell them.
Luskin here has continued in the tradition of the other chapters in this book by ignoring all of the best arguments that run contrary to his, while making previously refuted arguments with biased evidence, pretty much in line with what I
predicted before reading the chapter. He presents no positive case for a pervasively functional genome, and has only set out to cast doubt on the concept of junk DNA. Even in this, he has comprehensively failed. The book is called
Science and Human Origins, but the science is threadbare, and treated unevenly and unfairly.
*** UPDATE ***
A common theme in feeback I've been getting on Luskin's two chapters is that Luskin has been recycling his old material from years ago, which has all been soundly refuted.
One such case is with his argument against chromosomal fusion. In the past he has been much more outspoken about this than he was in the book - here Luskin has removed much of his previous chromosome fusion argument from the book, an unspoken admission that his arguments were without merit. But back in 2009 on Panda's Thumb, Dave Wisker had some fun with the chromome 2 fusion argument.
His four part series is a must-read, and there are valuable comments there as well (h/t Arthur Hunt).
24/7/12
Carl Zimmer had the cheek to ask for evidence about Luskin's chromosome 2 claims. The extraordinary non-response unfolded in a
Facebook thread. Absolute must-read stuff, especially Zimmer's linked content on his blog,
The Loom.
View comments