This is a post about the quality of science writing and what can be done about it. I'm picking on an article in SEED magazine here but it's not because SEED is any worse than the competition. It's partly because SEED makes claims about raising the quality of science writing and science education. For example, this statement from a SEED press release seems to indicate that they aspire to better science writing than the competition [Seed Media Group Adds Scientific and Political Pundits to Editorial Team] and certainly their commitment to science bloggers suggests the same aspiration.
As part of its growth strategy, Seed Media Group will develop original science content aimed at a general audience for distribution across a number of media channels, including magazines, books, newspapers, online, topical blogs, digital, film and television. Seed Media Group's endeavors will present science in the same culturally articulate and accessible style that earned Seed a prestigious UTNE Independent Press Award in 2004 and the support of leading advertisers.It's reasonable, in my opinion, to expect that SEED will live up to this billing. They've certainly made a major step in that direction by hiring PZ Myers to write a monthly science column. The science in his first two contributions is impeccable. As we will see, it raises the question of whether you need to be a scientist in order to get the science right. I hope that's not true.
I'm not going to criticize PZ's articles. Instead, I want to examine another article published in the March 2007 issue of SEED. (That's the one with the "TRUTH" prominently displayed on the cover!) The article in question is titled The Sound of Silence and it's written by Lindsay Bothwick, an experienced science writer with a M.Sc. from McGill and a Masters degree in journalism from Ryerson University here in Toronto. She's a senior editor at SEED so I'm assuming she can take criticism.
The article talks about silent mutations in protein-coding regions. The focus is on a recent Science paper showing that some silent mutations affect the activity of a protein. The point I will make is that the SEED article is very misleading and misrepresents the state of knowledge in this field.
Before getting into the article, let me give you some background.
The most common kinds of mutations are those where one nucleotide is substituted for another. For example, a G or an A or a T replaces a C. This substitution usually results from an error during DNA replication.
If the mutation (allele) persists in a population, it's called a single nucleotide polymorphism or SNP (pronounced snip). The term polymorphism means that there are at least two different alleles segregating in the population. Often these are the original "wild-type" allele and the new mutant allele.
We now recognize that genomes within a population are very heterogeneous. Polymorphism is common. This level of variation was discovered in studies during the 1960's and it's much higher than most scientists thought prior to 1960.
There are three explanations that can, in theory, account for this high level of polymorphism
First, if we think about SNP's, they can represent a transient phase of fixation by natural selection. In this case, one of the alleles is rapidly replacing the other and we just happen to catch it in the act. Back in the days when natural selection was the only game in town it was thought that this transient stage would be rare so populations were not expected to show much variation.
Polymorphism can also be explained by balancing selection. This is when the population has to maintain several different alleles because there is selection for heterogeneity. The classic example is the mutation for sickle cell anemia. When a person is homozygous for the mutant allele they exhibit the symptoms of anemia but when heterozygous they are resistant to malaria. Balancing selection is not common and it can't explain the variation that was discovered in the 1960's.
The third explanation was that the variation is mostly neutral. The idea here is that the majority of mutations are not being acted upon by natural selection. They are not being removed by purifying selection; they are not being maintained by balancing selection; and they are not rising to fixation under positive selection. It was the discovery of significant polymorphism in populations that gave rise to Neutral Theory in the first place ((Kimura, 1968, King and Jukes, 1969).
Neutral mutations will eventually become fixed or be eliminated from the population and the change in frequency is due entirely to random genetic drift. Drift is a much slower process than natural selection so there will always be large numbers of neutral alleles in the process of becoming fixed or extinct.
Neutral Theory and random genetic drift explains variation and it also explains molecular evolution and the (approximate) molecular clock. There are no other explanations that make sense and nobody has offered a competing explanation since Motoo Kimura (1968) or Jack King and Thomas Jukes (1969) published their papers almost fifty years ago. (Aside from occasional nitpicks, of course. There are always scientists who like to show that some mutations that were thought to be neutral are actually beneficial or deleterious. None of them have mounted a serious claim that most variation or most of molecular evolution can be explained by natural selection.)
The history of variation and the competing explanations were well covered by Lewontin in his 1974 book The Genetic Basis of Evolutionary Change. (Lewontin published the classic 1960's papers that revealed extensive within population variation.)
Long-held assumptions about "silent" genetic mutations have been torn down, challenging a fundamental evolutionary theory.
March 2007This brings me to the article in the March issue of SEED [The Sound of Silence]. It begins with a conclusion that's all too common in popular science writing these days,
Scattered throughout the human genome are thousands of mutations that biologists have treated mostly as footnotes. They're hardly few in number—in coding regions of the genome, there are as many as 15,000—but biologists regard them as mutations that simply don't change the way a cell functions. Both in name and effect, they have been accepted as "silent." Now, however, new discoveries are showing that silent mutations appear to play an important role in dozens of human genetic diseases, a fact that is forcing biologists to discard a long-held evolutionary theory and to reexamine the very rules governing the transfer of information from DNA to proteins.What's going on here? Has there been some extraordinary new discovery that's about to overthrow evolutionary theory and the "rules" of information flow? I will attempt to show that this rhetoric is completely unjustified. It presents a misleading picture of the state of modern science.
The author is talking about silent mutations. These are mutations in the coding region of a gene that alter a codon without changing the amino acid. The genetic code is redundant because there are 64 possible codons and only 20 amino acids. This means that several amino acids have multiple codons. For example, there are six codons for leucine (Leu): TTA, TTG, CTT, CTC, CTA, and CTG. If an original TTA codon is mutated to CTA then it still specifies leucine and this is a silent mutation.
The concept of codon bias has been known for almost forty years and it's an important part of all university courses in molecular biology. Some codons are more efficient than others during translation because the levels of various tRNAs in a cell are not identical. A rare codon will be translated less efficiently because the tRNA that binds to it will not be recognized as frequently as the codon for more abundant tRNAs. There are published codon usage tables for most species showing the preferred codons in that species. Highly expressed genes will preferentially use the codons recognized by the abundant tRNA species. All students know what these tables mean. It means that not all silent mutations are neutral. (It's on the exam!)
This is only one possible reason for silent mutations not being neutral. The various possibilites were discussed by Jukes back in 1980 when he revisited the evidence for neutral changes (Jukes, 1980)) . He gave some specific examples and then addressed the theoretical problem,
The question arises, are these silent changes actually neutral, or have they taken place for adaptive reasons, such as the requirement for a specific secondary structure in mRNA, or a preferential use for certain transfer RNAs in regulating the rate of synthesis of a protein?For some results the answer is that the silent changes really are neutral, although in a few cases there is evidence of adaptation. The point is that these issues have been recognized and dealt with for decades.
The existence of a few exceptions to a rule does not invalidate the generality. That's an important point. It's one that all science journalists need to grasp. There are no absolute, inviolate, rules in biology. The generalities are all about relative frequencies. Are most silent mutations neutral or are most subject to natural selection?
As Jukes put it 27 years ago,
The neutral approach to molecular evolution is a proposal to prove a negative, which is something like trying to show that a given substance is not a carcinogen. The counterrresponse to the publications by Kimura (1968) and King & Jukes (1969) has been quite strong. Any exceptions to neutrality are usually taken as disproof of it, and many authors have cited such exceptions for this purpose. We have, indeed, developed evidence for such exceptions ourselves, because a theory should be challenged by those who have postulated it.As the old expression goes, "those who are ignorant of history are doomed to repeat it." It helps a lot to be aware of the history of biology and the contributions of those who developed our current understanding. Modern science writers often fail to understand that there's not much that's new in biology these days. In this case, it's just not true that biologists were too stupid to recognize that some silent mutations weren't neutral. There was no "orthodoxy" that all silent mutations were neutral and, therefore, no orthodoxy has been overturned.
For example, the finding that synonymous codons for each amino acid are not use in equal amounts in β-hemoglobin mRNA has been cited as disproof of the neutral model, as if such a departure from randomness in a single gene were pertinent.
Silent mutations have no impact on the amino acid sequence of proteins and, therefore, were not expected to change their function.
March 2007The SEED article goes on to describe the results of experiments done by Kimchi-Safaty et al. (2007). They presented evidence that a cluster of three silent mutations in the MDR1 gene led to a slow down in translation and subsequent misfolding of the protein. Lindasy Bortwick then writes, "Through a series of elegant experiments, the team put to rest the idea that silent mutations were neutral." Of course, they did no such thing. They merely added one more data point to something that we already knew; namely, not all silent mutations are neutral.
Borthwick closes with,
Most fundamentally, the involvement of silent mutations in disease undermines the neutral theory of molecular evolution. This theory, posited by Motoo Kimura in the late 1960s and a powerful influence ever since, asserted that the vast majority of mutations were neutral, having no effect on the fitness of an organism, and spread through a population by chance. The fact that silent mutations are not harmless anomalies of nature means that they are not neutral. In contrast, some, if not all, silent sites must be subject to the forces of Darwinian natural selection.The theme of the article is that neutral theory is in big trouble. This point is emphasized in the highlighted quotations that are prominently displayed on page 35 (see the two boxes above). That's totally wrong and it distorts the modern consensus among knowledgeable scientists. Neutral Theory is alive and well, thank-you very much. It can easily accommodate one more example of a non-neutral mutation.
I believe that science writers have an obligation to get the concepts right and I believe they shouldn't misrepresent the science they're supposed to be presenting in a "culturally articulate and accessible style" to a general audience. A layperson reading this article would go away with the impression that a decades old concept has just been overthrown by a single paper published in Science. That's irresponsible journalism.
Jukes, T.H. (1980) Neutral Changes Revisited. In The Evolution of Protein Structure and Function, pp. 203-219.
Lewontin, R.C. (1974) The Paradox of Variation. in Evolution Mark Ridley ed., Oxfrod University Press, Oxford UK.
Kimchi-Sarfaty, C., Oh, J.M., Kim, I.W., Sauna, Z.E., Calcagno, A.M., Ambudkar, S.V., and Gottesman, M.M. (2007) A "silent" polymorphism in the MDR1 gene changes substrate specificity. Science 315, 525-528.
Kimura, M. (1968) Evolutionary rate at the molecular level. Nature 217, 624-626.
King,J.L. and Jukes,T.H. (1969) Non-Darwinian evolution. Science 164, 788-798.