The core dogma of conservation biology is clear: small populations are bad for species’ persistence. If we observe a population of endangered vertebrates harboring abundant deleterious mutations but without any reduction in fitness, what is happening there? I would like to bring up the curious case of the Channel Island foxes that was posted in the Molecular Ecologist blog a while ago and an interesting debate about using a small population as a source for genetic rescue that have been happening up until now.
To test the hypothesis that selection had purged strongly deleterious mutations in island foxes as the reason behind lack of inbreeding depression, Robinson et al (2018) simulated how the number of putatively deleterious mutations will change in six different demographic scenarios involving bottleneck in various periods. There are several interesting results from here, but what I find most interesting is that the island foxes end up having fewer strongly deleterious mutations than mainland foxes.
Small-sized populations. Having fewer deleterious mutations. Isn’t that kinda challenging the small-population dogma? That they’ll accumulate dangerous mutations if they keep being small in sizes due to genetic drift?
The bigger the badder?
A simulation study available as preprint by Kyriazis et al (2019) showed further that the larger the ancestral populations are before the bottleneck, the sooner the population will go extinct. These studies end up disfavouring genetic rescue from large outbred populations because of the potential of such populations containing more deleterious mutations than small populations that have experienced purging. They discuss how this will be different in populations with different rates of inbreeding though, but the general conclusion remains.
(If you have been following this debate, you surely know that this preprint has been updated and published in Evolutionary Letter a couple of months ago but please just bear with me for the sake of linear storyline)
Since it was out in the wild world of the internet, Kyriazis et al (2019) preprint has been used by some people to discuss how the small population paradigm has not been the fate for all endangered populations. After observing the literature on this topic for quite a while, there actually exist species that have been persisting as very small populations for quite some time and then bounced back just fine!
However, I noticed also that most persist because of humans’ intervention, be that strict protection of habitat, a hunting ban, or adding congeneric species to assist their breeding (note that I use “congeneric”, meaning that it is even not the same species/subspecies/lineage). We should not forget about other cases where inbreeding within small-sized populations ends up having a really low fitness without any signs of purging like the Chatham Island black robin (Kennedy et al 2014).
Noticing that this has been challenging a very core practice of conservation genetics, there must be something we have missed. What causes the high number of deleterious mutations in large populations? Does this mean adding individuals from large-sized populations is bad? Everyone has so many questions.
Here comes the critics
Worrying that policy-makers and practitioners will have more scientific evidence to back their reluctance to do species translocation, Ralls et al (2020) published a thorough critique of Robinson et al (2019) and Kyriazis et al (2019) approach. They wrote five sections explaining their arguments, some of them have subsections, and I would like to sum them differently:
Critic #1 Power of the study
Considering the many empirical evidence of inbreeding depression coming from populations with low genomic diversity, the Channel Island foxes can be seen as survivor bias. Their assessment of inbreeding depression also needs to include traits that directly related to fitness such as litter size and island-specific disease susceptibility. Empirical evidence around the benefit of using a large outbred population as source for genetic rescue are also abundant, so that recommendations defying this trend need a stronger ground.
Critic #2 Definition of “genetic rescue”
Robinson et al (2019) bring up the case of the Isle Royale wolves as an example where a large outbred population contributed migrants into a small population, and made the small population decline even further. Ralls et al (2020) stated that the case “does not represent a good model for a well-conducted genetic rescue”. Of course if the populations end up declining, we cannot really say that it is a “rescue”!
Both Ralls et al (2020) and Robinson et al (2019) here acknowledge the importance of considering population history in choosing population source in this case. The difference is that Ralls et al (2020) consider the population history of the recipient Isle Royale before the migrant came while Robinson et al (2019) and Kyriazis et al (2019) focuses more on the population history where the migrant came from.
Critic #3 Purged populations cannot be fully trusted
To sum it up, purging is not effective (1) in variations that is not strongly harmful, (2) if caused by drift rather than actual selection or non-random mating, (3) in all harmful alleles in small isolated populations, (4) in all loci subject to balancing selection, and (5) in environments other than where the purging happened. Also, when purging is effective, it reduces genetic variation elsewhere in the genome and further reduces adaptive potential.
Critic #4 Unrealistic key parameters
Ralls et al (2020) also argue that the parameters used in the simulation do not match theory and data from real populations. All loci contributing to inbreeding depression are modelled as fully recessive. Empirical studies do not agree with this. A lot of harmful mutations are actually only partially recessive, not fully, and thus behave very differently than fully recessive ones. This also further makes such mutations less sensitive to purging.
Another relevant things that are more difficult to include are beneficial mutations, variation in non-coding regions, and possible adaptive potential after genetic rescue. Therefore, any conclusion should be made with the realisation that these factors are not considered.
How it all ends (or continues?)
After Ralls et al (2020) paper, the preprint got updated shortly after. The new Kyriazis et al (2020) preprint did not only follow the suggestion to use different dominance and selection coefficient but also change their title and main text to emphasize the deleterious mutations as the problem rather than the genetic rescue approach.
The results do not change, however, and large populations that just recently bottlenecked have more deleterious mutations than smaller populations. Possibly because they still do not model beneficial mutations along with these weakly deleterious mutations, which is relatively hard to do. A major update nonetheless is the various dominance coefficients involved in the simulation following Ralls et al (2020) critics. They also differentiate this between populations before and after bottleneck.
Simulating adaptive potential is hard though; how do you decide the direction of evolvability? I have high hopes for SLiM and the many new genomic simulation software nowadays though (please tell me if you are working on this!). From the PopGen Group Meeting, I also learned quite a number of works going on about understanding direction of evolution in the genetic level. Anyways, bottlenecked populations having more deleterious mutations are not unreasonable, as was recently found in the Alpine ibex.
Despite the limitations and controversy, I still root for the conclusion of Kyriazis et al (2019) that “where populations are destined to remain small and isolated, management strategies should aim to minimize deleterious variation rather than maximize genetic diversity”. I would like to understand that part as referring to places with limited funds available to maintain an insurance population to avoid the species’ extinction. In such cases, it is often easier to just focusing more on providing more habitat and resources rather than focusing on getting a new genetic variation, allowing the population to grow healthier than they have been. This part is rephrased in the latest version in the Evolution Letter (just published in December 2020, please have a read yourself!) as they conclude how keeping the existing populations healthy is more important than adding new individuals.
The Isle Royale wolves still make a good case where we need to be concerned with other factors than the genetic makeup of migrant and recipient populations, and I think the Florida panther case that were referred as “a better model of the probable results of a genetic rescue” by Ralls et al (2020) should be acknowledged as a success exactly because they were considering environmental factor such as habitat availability in addition to genetic variation. If habitat quality and quantity can be guaranteed in the long term, a small population with trustworthy demographic history can help other small populations avoid extinction. Our ability to simulate a more realistic scenario with genetic and environmental data will help us understand more… or not.