Relatedness and kin discrimination in vampire bats (and a few updates)

It’s crucial for my future experiments to have very precise estimates of relatedness of “my” vampire bats. While working on other things (like scoring videos), I’ve been slowly adding microsatellite markers to increase the precision of my measures of relatedness. (By the way, one of the best explanations of what relatedness is, can be found here).

I have about 25 polymorphic markers now, but hope to have 30-40 by the end of the summer. Rob Fleischer, head of Smithsonian Center for Conservation and Evolutionary Genetics in DC is helping me get sequence data using a Roche 454 Jr.

It’s comforting when bats with known pedigree kinship values (like mothers and daughters) converge on the correct relatedness values of 0.50 as you add more markers. But it’s also interesting to look for related pairs in groups that I have not worked with much. Sometimes when you catch vampire bats in a mist net around cattle, you’ll get up to 15 all in the net at once, as if they were flying like a group. I’ve always wondered if these bats are relatives (like daughters following their mothers). Yet, after genotyping bats caught in mist nets all together, I have not noticed any evidence for this. Most are just as genetically similar to each other as bats caught at completely different sites.

The new relatedness estimates I have further confirm the result I have already published: social experience seems to affect food-sharing decisions more than genetic relatedness. With the new relatedness estimates (calculated using ML-Relate) and using the relaimpo R package, I calculated new estimates of relative importance of several factors in my updated model that explains 38% of the variation in amount of food sharing. Yeah I know, there’s a lot of residual variation. My excuses are: social behavior is complicated, I did not have a measure of how hungry the bats were (obviously important), and donation rates were estimated from mouth-licking times (correlated with weight gains but still a noisy measure of food sharing).

The amount of reciprocal help from the partner explains 50% of the explained variance, the donor’s sex explains an extra 24%, and the social grooming rate with the partner explains 19%. But the partner’s genetic relatedness explains only 7% and is not significant in the model.

Note: This does not mean that relatedness is unimportant. There are at least two other interpretations I think about often:

First, relatedness might be quite important, but just not as important as the cooperative social bonds between bats. In the captive population I last studied, relatives often fed each other in both directions (supporting both reciprocal help and kinship), but so did several unrelated pairs of bats, which is one reason why reciprocal help explained food sharing more than relatedness. In a previous study, I failed to find that flower bats could learn a simple shape-based learning task, when spatial cues were available. But this wasn’t because they couldn’t do it, but the effect of shape was so overwhelmed by spatial memory that demonstrating shape-based learning would have required about 5,000 more trials. My point is that social experience might simply and similarly overshadow kinship cues.

Second, in vampire bats, kin discrimination may rely largely on familiarity. For example, humans treat family members more favorably than strangers, but this preference is mostly based on earlier social experience and not on direct assessments phenotype similarity (although there is evidence for that too in humans). Probably most forms of kin discrimination involve learning of some kind, except obviously in organisms that lack cognition but still recognize kin, like plants.

In the early days of inclusive fitness theory, many people assumed that kin discrimination would occur mainly by phenotype matching, say, by one animal smelling another’s scent and then comparing this to its own scent. Richard Dawkins called this the “armpit effect“. There is now good evidence for this self-matching mechanism for kin discrimination in birds, mammals, and fish.

However, kin discrimination may also be based largely or completely on prior association, or on templates from other family members rather than one’s self.

Interestingly, kin recognition based on simple phenotypic markers is not evolutionarily stable and drives itself extinct. This was first realized in a paper with a great title: “Genetic clonal recognition abilities in marine invertebrates must be maintained by selection for something else” by R. Crozier. Very briefly, the basic theoretical problem is that if organisms use any specific marker to identify relatives as targets of altruism, then this extremely useful marker tends to become so successful and so common that it becomes fixed across the whole population. At this point, it is useless as a marker of genetic relatedness, because it does not give the individual an advantage over others in the population. It eventually loses its evolutionary value and becomes unlinked to altruism.

This problem is the same with so-called greenbeard genes (another term coined by Dawkins), which are genes that simultaneously encode both a trait and a tendency to help those with that trait. For example, imagine a gene that gives you a green beard and makes you nice to people with green beards. Again, Dawkins suggested this as a thought-experiment, and again it turned out to be real (the basic mechanism, not beards that are green).

As Crozier suggested, “something else” must be maintaining these markers for genetic kin recognition. One solution is if some mysterious force somehow drives genetic markers to be rare (and hence informative of true kinship)– i.e, balancing selection.

This mysterious force appears to be… parasites. It turns out that when animals smell each other, they are often matching the so-called MHC genes (for major histocompatability genes), which are involved in the immune system. MHC genes are “smelled” by fish, rodents and humans. These genes are very diverse, because they are used for defense against rapidly evolving pathogens/parasites. Hence, so far parasites and pathogens to be what create the necessary selective pressures for the existence of both true kin recognition and sex*

*males (which don’t even reproduce) seem to have evolved as a tool for environmental sensing and updating, allowing females to stay ahead of natural selection and make offspring that were adapted to the parasites of today rather than yesterday. By inventing males, females could recombine genes with those that survived under the current parasite conditions– just toss ’em out there and reel back the healthy ones.

But let’s get back to kin recognition. In some cases, animals just use simple cues for kinship, like prior association. In other cases, self-matching can be used when association cues might fail. For example, after their first winter, young Belding’s ground squirrels recognize their true siblings, but no longer identify with “adopted siblings” with whom they had been raised since birth. Why? Because while in hibernation they forgot their kin recognition templates, and had to use “the armpit effect” again to determine who are their relatives.

In an interesting contrast, hibernating bats seem to remember their past associates even after awaking from their deep slumber. Bechstein’s bats maintain long-term relationships over years, even though social groups completely disappear over the winter and are reconstituted the next summer (like kids at a summer camp).

So kin discrimination is clearly a complicated affair that can involve interactions between prior association and phenotype matching.

And we still don’t know how vampire bats discriminate their kin. They do seem to be at least able to discriminate kin, because they favor helping relatives in the wild even when controlling for association. I’ve discussed elsewhere why I think vampire bats don’t just use familiarity to identify kin. But it would be good to figure out how they do it.

As a first step, I’m seeing now if genetic similarity maps to similarity of their social calls. It would be quite interesting if vampire bats can identify relatives at a distance (rather than just sniffing them), because this would allow even kin (even unfamiliar kin?) to end up in the same roosts together when switching roosts frequently.

Some other recent updates

• my previous work was highlighted in Current Biology.
• received some awards and small grants, including a grant from the Animal Behavior Society, a travel award from the Biology Department, the annual research award from the BEES graduate program, and a NSF DDIG.
• Recently have a book chapter in press (Cooperation and Conflict in the Social Lives of Bats) and submitted an invited review (Does food sharing in vampire bats demonstrate reciprocity?). Send me an email for a manuscript copy of either.
• I came across this truly amazing online tutorial on the evolution of cooperation.
• Next week, I’m talking at the Mammal Meetings, then the next day, heading out to Trinidad to help my labmate Danielle with her project on male reproductive success in greater spear-nosed bats. Basically, we’ll be counting baby bats in cave. That’s a pretty great vacation as I see it!