Birch and Bentley: Time and relatedness in microbes and humans

Dr Jonathan Birch from the LSE is working on a book called 'The Philosophy of Social Evolution'. For this meeting Birch drew on his recent paper 'Gene mobility and the concept of relatedness' to talk about a foundational idea in contemporary evolutionary theory - Hamilton's theory of Kin Selection - and in particular at its application in the context of what has been called 'sociomicrobiology' - the study of sociality in bacteria and other microbes.

Cooperation in the microbial world is complicated, Birch claims, by the phenomenon of gene mobility. Genes are mobile when they participate in mechanisms of lateral gene transfer. There are three different mechanisms for transfer between bacteria. Cells can take up free-floating dna from their environment and incorporate it into their own ('transformation'). They can swap dna with each other via special tubes ('conjugation'). And they receive packets of dna carried by viral vectors known as bacteriophages ('transduction'). In each case, the cell ends up with genes in its genome that didn't come from its parent cell. So we say the genes are propagating laterally, instead of horizontally, via these mechanisms.

Birch argues that in order to accommodate lateral gene transfer within Kin Selection Theory (see Clarke's Introduction), we need to radically reconceive the central concept within that theory: relatedness, often written as r. Rather than assuming that r is a life-time property of an organism, as is standard, Birch claims that we must consider relatedness as changing over time. To be precise, we must use a diachronic measure, on which what matters is the extent to which there is genetic similarity between a donor cell at the time it manifests an altruistic behaviour, and a recipient cell at the time that it reproduces, where these two times need not be contemporaneous. To put it another way, genotypes need to be indexed to time.

This reconceptualisation is necessary to accommodate the possibility of selection for something Birch calls 'Ship-jumping'. This would occur when a cell exhibits altruism towards unrelated, non-altruists, in the expectation, thanks to lateral gene transfer, that the recipients will become relatives in the future. Bacterial cells would be investing in strangers, only to subsequently harvest the returns of their investment by converting the strangers into relatives, by directly transferring their dna to them. This altruistic investment could give a laterally transferring gene a considerable advantage, over and above the advantage it gains from mere infectivity alone.

It is an elegant proposal.  And Birch points out the radical implication that an altruistic trait could be selected for in the absence of correlated interaction between the bearers of the gene for the trait. At least, in the absence of correlated interaction as we normally think of it: in the moment the altruistic act is carried out, the recipients of the act are neither relatives of the donor, nor do they need to be altruists themselves.

How far is the idea of 'ship-jumping' grounded in reality? Well, we know that lateral transfer takes place at an increased rate in the densely packed conditions of microbial biofilms. The scale at which cooperative interactions take place, via diffusible public goods, for instance, is likely to be similar to the scale at which lateral transfer occurs. So if a cell were to dish out favours to its local community, it can be pretty confident in being able to access that same market as lateral gene recipients, assuming it has some means for transferring genes to them. There isn't too much danger of the nearby cells taking the money and running, in other words. A biofilm creates a fairly captive audience.

What is a bit murkier, however, is the extent to which bacterial cells have a say over who they do or don't swap genes with. In the case of transduction, for example, it seems like gene mobility is all under the control of the phage - the bacterial cells are just used by the phage who are seeking to maximise phage fitness. This doesn't undermine the idea that a phage might use the strategy of fattening its host cells up before infecting them. Indeed, it is now firmly established that cooperation is over-represented on mobile genes (See Dimitriu et al 2014). But the 'selfish phage' interpretation does play around with the idea that laterally transferred public good production is correctly understood as bacterial cooperation. The phage tokens are cooperating with one another. Or perhaps they are better thought of as exhibiting a sort of parental care....preparing a cosy nest for their babies. But any benefits to the bacteria are just a serendipitous side-effect. Crucially, it need not be the case that the trait has been selected in virtue of its effects upon the recipient bacterial cells. Rather, it may have been selected in virtue of its effects upon the phage.

With transformation and conjugation, there is more scope for the cells themselves to wield control over when and whether they pick up new dna. For the ship-jumping or 'fatten-and-then-recruit' strategy to be successful, the actor must retain some confidence that recruitment will be a success. The best way to do this, presumably, would be to do things the other way around: recruit and only then pay out. So perhaps we can expect that most of the time, altruism will be directed at already-kin. Nonetheless, Birch succeeds in dissecting the notion of relatedness and demonstrating that the relationship between the donors and recipients of altruism may be much more complicated than the standard concept of relatedness implies.

Birch went on to think about possible comparisons between bacterial cultures and human cultures. He noted that one very tempting idea is to think of a property that is 'cultural relatedness' as helping to explain cooperation between humans. He notes that microbial work might be helpful in understanding how cultural relatedness could work, because both microbes and humans engage in a lot of horizontal transfer. In microbes this is via gene mobility, of course, whilst in humans its mediated by social learning. Birch suggested that his insights about Hamilton's rule might apply to culture also, so that it might be necessary to understand cultural relatedness as a diachronic property.

Birch asked if cultural ship-jumping takes place? Professors invest in their graduate students, hoping that the students will ultimately uphold their legacy. An audience member suggested the example of unscrupulous religious groups going out dispensing food to the needy in the hope of converting them, for another example. Finally, Birch pointed out further analogs between microbial and cultural cooperation consist in the vexing questions that persist, in each case, about why altruists are not suppressed by the rest of the genome/culturome, given that levels of relatedness will be very variable across different parts of those genomes/culturomes.

Dr Birch received a commentary from Michael Bentley who is close to completing his PhD in microbiology. Bentley works on time scales and on the issues that time scales can create for the dynamics of systems that operate on multiple hierarchical levels. Bentley was all in favour of seeking  mathematical frameworks that can be universalised to suit model systems as different as microbes from the multicelular organisms for which Hamilton initially developed Kin Selection Theory, and from humans. He would prefer an even more general measure of relatedness: in particular one that escapes being limited to genetics, and incorporates all causes of additive genetic value. In other words, he would prefer a measure of relatedness that accommodates extra-genetic sources of similarity between parents and offspring, such as methylation patterns.

Bentley noted further that there are various alternative perspectives on lateral gene transfer. Rather than think about a plasmid as transferring form one individual to another, we might instead think about a non-plasmid-bearing individual dying at the moment that a new plasmid-bearing individual is born. Or we could think about a population of cells in which frequency-dependent mutation causes indivdiuals to become cooperative. It all depends on how you understand core concepts such as 'reproduction', and leads to differing mathematical models of the dynamics. The point is that there are multiple units and levels of selections in play. Birch chose to view things from one level. But how important is this choice? Bentley worries that Hamilton's Rule obscures too much of this important detail.

You can listen to Dr Birch's talk here. Michael Bentley's response begins 56 mins 52 secs in.

Next post in series: Brown and Heyes - Social learning and the other cooperation problem.