A recent Science Daily article blares: “New Findings Confirm Darwin’s Theory: Evolution Not Random” But while the research looks legitimate, and the findings indeed consistent with evolution via natural selection, the article is deeply confusing, and it looks like we have another example of shoddy science journalism on our hands.
Right off the bat, the article claims that “An opposing theory [to natural selection] says evolution takes place through randomly inherited and not necessarily advantageous changes.” What this opposing theory is, however, goes unexplained. Casual readers might well assume that the article is talking about creationist claims, since creationists are constantly harping on the idea that evolution is a “random” process (and thus cannot explain anything).
My guess is that what the researchers were actually comparing was the explanatory power of the neutral theory against that of natural selection. The neutral theory, most commonly associated with Japanese biologist Motoo Kimura, is basically the idea that the vast majority of changes to DNA over time do not have any particular affect on fitness or even morphology (i.e. the actual outward traits of the species). They are, as it were, just the result of chaotic genetic drifting over time.
If this is what the article is talking about, then it’s being extremely misleading. It sets up the experimental dispute as either-or situation, and then declares a definitive win for natural selection. However, most biologists agree that both neutral changes and genetic drift AND natural selection both play a role in biological life. The question is just to what degree and in what context these effects show up or dominate each other, not a matter of disproving one in favor of the other.
The two ideas aren’t even precisely at odds even in the abstract. Neutral theory, for instance, is primarily a base-pair level view of a creatures genome: showing that if we watch the DNA of a gene pool over time, the changes to base-pairs that are retained or tossed out will be almost entirely statistically random. This does not, however, really contradict natural selection: traits that are selected for could all well fit comfortably in the realm of that “almost entirely” exception. Furthermore, traditional natural selection acts on expressed traits (i.e. the phenotype) rather than on genes directly (though there are actually some gene-level selection gradients at play as well, such as sequences that code for reproducing more copies of themselves). In some ways it’s really a different level of observation entirely.
You can also think of neutral theory as a sort of “null hypothesis” when looking at change in genetic sequences over time: if you see some sequences vanish and others appear, the basic presumption is that this is NOT, in fact, proof of natural selection at work. But the point of this view is not to assert that natural selection happens: on the contrary, it’s to help give a sense of the general background noise over which natural selection can be observed.
So the issue under experiment in this case is actually far less dramatic and far more of a complex technical subtly than it might seem to unwitting readers. Either the unnamed author of this article, or the researchers themselves (the article claims to have been “adapted from materials provided by the American Technion Society) needs a nice slap on the wrist for shoddy science journalism.
Update: I finally tracked down a link to the actual paper. As I suspected, and the paper speaks for itself far better than the science journalist did:
A surprising amount of developmental variation has been observed for otherwise highly conserved features, a phenomenon known as developmental system drift. Either stochastic processes (e.g., drift and absence of selection-independent constraints) or deterministic processes (e.g., selection or constraints) could be the predominate mechanism for the evolution of such variation. We tested whether evolutionary patterns of change were unbiased or biased, as predicted by the stochastic or deterministic hypotheses, respectively.
We propose that developmental evolution is primarily governed by selection and/or selection-independent constraints, not stochastic processes such as drift in unconstrained phenotypic space.
So, an interesting result as far as understanding developmental genetics, but as to any sort of radical new proof of natural selection over “random” evolution, or even an attempt at such, nothing to see here. Other than lots of worm vulva. Ewww.
Double Super Update: The original article was worse than I thought. According to the actual paper, the researchers plotted a phylogeny based on the observation of extant species and then analyzed their genetic and trait differences that way to figure out whether there was evidence of selection bias. Much like the RNA researchers I discussed a little while ago. But the article, primarily via omission, leaves the impression that the researchers actually watched the nematode vulvas evolve over time, “measuring changes” as they happened. What a mess!