26 October 2017 by pencechp • permalink
In a footnote to a great paper on bacterial systematics by Laura Franklin-Hall, my attention was drawn to the following aside that John Maynard Smith offers in one of his papers on the same subject. After wondering whether it makes sense to think of bacterial populations as composed of groups of reproductively isolated clones, he muses:
At the opposite extreme, is it better to take a wholly gene-centered view of bacterial evolution, and regard the bacterial cell – or, rather, the bacterial chromosome – as merely a temporary alliance of genes, analogous to a European football team, composed of players from many different countries, all liable to be transferred at any time?1
Now, if you know me well, you know that one of the only passions I have that matches my appreciation of the philosophy and history of biology is my passion for The Beautiful Game. So naturally, as soon as I saw this metaphor, I knew that I would have to take a stab at running it completely into the ground. This post, then, will be an exercise in pushing a metaphor far, far past its breaking point.
First, the basics on the biological side. From the perspective of sexual, animal mating like we’re used to, bacteria are entirely strange creatures. They engage in massive, constant genetic transfer in their natural environments – swapping genes around regularly, even among relatively distantly related species. We call this horizontal or lateral transfer, to indicate that it’s not happening “vertically,” at the time when the organisms give birth to their offspring, but rather at arbitrary moments during the life of an organism. You might imagine going to the park to watch your kids, and coming home with a different eye color. And this horizontal transfer has been responsible for the spread of some really significant genes, such as the ability of soil bacteria to fix nitrogen, as well as resistance to toxic materials and antibiotics.2
From the bacteria’s perspective, on the other hand, we are highly promiscuous in the extent to which we swap genetic material. Bacteria only trade genetic material in isolated, small bits, unlike we profligate creatures, who mix fully half of our DNA with that of a mate to produce children.
Why is this evolutionarily interesting, and what does it have to do with a football team? Well, for individual humans, it’s very easy to draw up a lineage. All of your DNA came from your two parents, and all of their DNA from their two parents, and so on back for millions of years.3 That makes it a simple matter to pick out who counts as humans and who doesn’t (at least until we get back to crosses between anatomically modern humans and other species, which hasn’t been happening for a while now).
But here’s where the football team metaphor comes in. If horizontal transfer is really common in bacteria, then it may well be the case that each individual gene that a bacterium carries around has a different pedigree. The genes, that is, start to look like independent agents themselves, untethered from the requirement that they be passed from parents to offspring at the moment of bacterial cell division. And a bacterium might kick out a gene and send it along to its neighbors at any moment. Now, hopefully, you’re starting to see the metaphor. Genes aren’t under long-term contracts with their teams like they are in animals – rather, they could get transferred by the manager at any moment. And the characteristics of a bacterium – even some of the more important ones, like ability to fix nitrogen or resist a toxin – are not indelible features of that bacterial species, but rather temporary properties that result from who’s under contract at the moment. (What this might mean for an “essentialist” reading of bacterial species – or species in general – is a project for another day.)
Now, let’s take this metaphor from the reasonable to the ridiculous. What other parts of the analogy might we explore? The first significant disanalogy worth pointing out is that the players in the bacterial case aren’t unique objects – genes can, of course, be copied. So if Antoine Griezmann has a good season, copies of him can proliferate and find themselves in leagues all over the world. Similarly, since bacteria form populations, “teams” aren’t really unique either – imagine a “population” of ten million versions of Manchester United, each of which vary ever so slightly in their fourth-string bench players (*shudder*).
How about the actual football matches? Success and failure, at least on a grand scale, comes down in many cases to the standard evolutionary forces of natural selection and genetic drift. And another facet of the metaphor successfully tracks here, as well – individual players don’t usually compete, except as members of a team.
What about the managers and technical directors and directors of football – who’s deciding who gets transferred and who gets the ax? Here we have to point to a whole cluster of biochemical processes, many of which are not often enough discussed in our high-level treatments of evolutionary theory. (Aha! Our metaphor is producing some dividends!) A paper by Peter Sneath points out three main mechanisms: conjugation, transformation, and transduction. Conjugation is a bit like sexual reproduction, with two bacterial cells coming together and transferring genetic material via a physical connection between the two. Transformation involves the taking up of free-floating DNA from the surrounding environment. Transduction involves the mediation of bacterial viruses, known as bacteriophages – a phage picks up some genetic material from bacterium A, and passes it on to B when it becomes infected. (That last method is, however rare, one of the methods by which horizontal transfer might occasionally happen in eukaryotes.) And these aren’t just interesting details of the DNA replication machinery, relevant only if you care about the nuts and bolts – they’re determining when pieces of DNA are being horizontally transferred and when they aren’t! (As an aside, the operation of this machinery is likely to have some interesting upshot for the role of chance in evolution as well, a topic that I hope to consider more at some point in the future.)
I think we can extract one more piece of payoff before this metaphor completely disintegrates under the strain. A significant amount of analytical energy has been expended in attempting to produce “expected goals” (xG) models for soccer matches (there’s an example at the left). The idea here is that soccer is a wide-open, highly chancy game, subject to all sorts of random events, clearances off the line, shots into the woodwork, passes off by a centimeter, and so forth. (Anyone who has watched their favored side crash attempt after attempt into a defense that somehow manages to keep the ball out of the net knows exactly what I mean.) The xG models are supposed to “cut through” all that – by telling you how the team should have done, in some strong sense, given the quality of the shots that they managed to create. If, that is, a team took that same set of shots a thousand times, how many goals should they have managed to net?
We have here, I think, an interesting analogy with the determination of fitness in evolution. Of course, organisms will very often (almost always) fail to perform in the way that they “should” given the expectations about their genotype in a given environment. But developing those models, and comparing them with actual organismic outcomes, still tells us something interesting about what’s happening in a given evolutionary scenario. Fitness, like xG, is still a quantity worth measuring and trying to understand, even if its actual impact on the history of life is hampered by the impact of chance.4
I don’t expect everyone to rush out and rewrite their textbooks to include pictures of Messi. But any time I can use a fun metaphor to think about some interesting questions in evolution, I’m game.
At the bottom of the first page of the 1990 article “The Evolution of Prokaryotes: Does Sex Matter?” This article is unfortunately paywalled, though it can fairly easily be found with a quick search. ↩
It’s a bit empirically unclear at the moment just how common HGT really is. The relevant research is hard to do, and the data are unsettled. But it assuredly occurs fairly often, and some studies place the fraction of horizontally transferred genes as high as a quarter. One author estimated the rate of lateral transfer at 4.4x that of mutation. ↩