One of the many fundamental insights to come out of the early days of population genetics in the first decades of the 20th Century was J.B.S. Haldane’s discovery that, when it comes to natural selection, population size matters. As Haldane laid out in a 1927 paper for the Cambridge Philosophical Society, the probability that a new, beneficial, dominant mutation will not be lost to the “mere bad luck” of genetic drift is approximately two times the relative advantage it confers—and it is more likely to persist from generation to generation as the number of individuals carrying it increases.
A paper published last year in Nature Communications reported a direct observation of this relationship between the number of individuals carrying a mutation and its chances of long-term survival. Ivo Chelo and colleagues created artificial populations of inbred, highly homozygous hermaphroditic Caenorhabiditis elegant nematodes from two inbred lines with measurable fitness differences and one line, the fitter one, engineered to produce green fluorescent protein.
The authors could then introduce different numbers of GFP-tagged worms into replicate populations of the less-fit non-GFP-tagged worms, and track the fate of the fitter line by counting the green-glowing worms in each population. Indeed, when they introduced five GFP worms, the GFP lines were more likely to persist and spread than when they started with only two.
That’s a remarkably concrete demonstration of Haldane’s original result, though it’s hard to know whether such a tightly controlled experiment is truly better than a computer simulation—the authors even enforced discrete generations on their experimental worm populations. Can we really expect to see results very different from Haldane’s model from such a carefully controlled, if biological, system?
Well, maybe? I’d like to think that “An experimental test on the probability of extinction of new genetic variants” is as much a test run of a new experimental system as it is intended to validate one of the foundational results of population genetics—and we’ll soon see the authors coaxing some much more complex dynamics from their green fluorescent nematodes.
The other result associated with the Haldane (1927) citation is that, all else being equal, a beneficial mutation is much more likely to gain selective traction if its effect is perceptible in individuals carrying only one copy—that is, when the mutant allele is dominant. This is “Haldane’s sieve,” and in addition to being a classic of population genetics, it’s not a bad metaphor for the process of science itself. Which is to say, what we choose to investigate, and how we build explanations, depend on what we notice.
To take a personal example: a lot of my dissertation research is about the pollination biology of Joshua trees. Specifically, trying to decide whether their pollinators are exerting divergent selection on (or otherwise responsible for) the shape of the trees’ flowers. Joshua trees have no other functional pollinators except for yucca moths, which lay their eggs in Joshua tree flowers and use specialized mothparts to deliver pollen with a remarkable degree of efficiency and deliberateness, to the extent that a tiny black moth can do anything deliberately.
If there is a plant-pollinator relationship in which a single pollinator probably has a selective effect on the plants it visits, this is it. The probable selective impact of Joshua tree’s pollinators is obvious—it easily catches in the sieve of our attention and our desire to work with interesting critters. But I think it’s also fair to ask how much an interaction as specialized as the Joshua tree pollination mutualism actually tells us about the evolution of much more common, much less finely co-adapted relationships.
The tendency to get to the obvious cases first may not positively mislead us, but it does have an effect on how science proceeds. An example that springs to mind turned up in an 2010 article for the New Yorker that occasioned eye-rolling from most working scientists, by Jonah Lehrer (he of the manufactured Bob Dylan quotes, self- and nonself-plagiarism, and general loose attitude towards actual facts). Lehrer described a so-called “decline effect,” in which scientific results become weaker over time, and repeated attempts to replicate them.
One of Lehrer’s case studies for the decline effect was the history of “fluctuating asymmetry,” in which sexual selection manifests as a (usually) female preference for more symmetrical (usually) males—the idea being that greater symmetry indicates more perfect development, and is a proxy for greater genetic and conditional fitness. This effect was first documented in a study showing that male barn swallows with longer, more symmetrical tails had more mating success. Barn swallows seem, in retrospect, like a pretty good case in which to notice an effect of female preference on male ornamentation. Subsequent studies found similar effects in other species, as Lehrer describes it, but as the years progressed more and more studies in new species looked for fluctuating asymmetry and failed to find it.
This may be an effect of publication bias, journals eager to publish results documenting an interesting new phenomenon, but not so much in negative findings until the trendy new buzzword is old hat and it becomes more interesting to start showing when it doesn’t apply. But I think it also has something to do with the sieve of scientific attention.
Scientists come up with new hypotheses when they’re struck by something that obviously needs explaining, in whatever critter or system they study. It shouldn’t be surprising that the first documentation of a particular physical or biological process would be one of the strongest—and then, as other scientists read about the new concept and think about whether it applies in their own work, a round of cases emerge in which the new concept turns out to be a pretty good fit. But as time passes, the concept becomes well known, and more and more authors invoke when they’ve found a pattern that could be consistent, but maybe isn’t a perfectly satisfactory fit.
And this isn’t necessarily a problem! But it is something to keep in mind, as we survey the scientific literature to try and build a collective understanding of processes that have been studied over and over by different authors, in different labs, using different models and study systems. What we can find out, with our favorite versions of glow-in-the-dark worms, may be limited by the very things that make them interesting in the first place.
References
Chelo IM, J Nédli, I Gordo, and H Teotónio. 2013. An experimental test on the probability of extinction of new genetic variants. Nature Communications. 4:2417. doi: 10.1038/ncomms3417.
Haldane JBS, 1927. A mathematical theory of natural and artificial selection, part V: Selection and mutation. Proc. Camb. Philol. Soc. 23:838–844. doi: 10.1017/S0305004100015644.
Møller AP. 1992. Female swallow preference for symmetrical male sexual ornaments. Nature. 357:238-240 doi:10.1038/357238a0.
Yoder JB, CI Smith, DJ Rowley, WKW Godsoe, CS Drummond, and O Pellmyr. 2013. Gene flow in Joshua tree (Yucca brevifolia) populations shaped by the consequences of pollinator divergence. Journal of Evolutionary Biology. 26: 1220-1233. doi: 10.1111/jeb.12134.
