Living with relatives offsets the harm caused by pathogens in natural populations

  1. Hanna M Bensch
  2. Emily A O'Connor
  3. Charlie Kinahan Cornwallis

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    Group living may be beneficial for many reasons, but has costs in terms of increased rates of parasitism, in particular if group members are highly related. In this meta analysis, many original studies on questions related to parasitism, relatedness and group living are brought together in one unifying framework. The authors conclude that living in groups can indeed facilitate the spread of infectious diseases, but that these costs can be overcompensated by the benefits of group living.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #2 agreed to share their name with the authors.)

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Living with relatives can be highly beneficial, enhancing reproduction and survival. High relatedness can, however, increase susceptibility to pathogens. Here, we examine whether the benefits of living with relatives offset the harm caused by pathogens, and if this depends on whether species typically live with kin. Using comparative meta-analysis of plants, animals, and a bacterium ( n species = 56), we show that high within-group relatedness increases mortality when pathogens are present. In contrast, mortality decreased with relatedness when pathogens were rare, particularly in species that live with kin. Furthermore, across groups variation in mortality was lower when relatedness was high, but abundances of pathogens were more variable. The effects of within-group relatedness were only evident when pathogens were experimentally manipulated, suggesting that the harm caused by pathogens is masked by the benefits of living with relatives in nature. These results highlight the importance of kin selection for understanding disease spread in natural populations.

Article activity feed

  1. Version published to 10.7554/elife.66649 on eLife
  2. eLife

    Author Response:

    Reviewer #1:

    This meta analysis addresses a double-edged sword in evolutionary biology. Group living may be beneficial for many reasons, but has costs in terms of increased rates of parasitism. Furthermore, if groups are highly related, parasites that are genetically able to infect on member of the group may be able to infect all of them, putting the entire group at risk. In the her presented meta analysis, many original studies working on questions related to parasitism, relatedness and group living are brought together in one unifying framework. The authors find that indeed, group living can facilitate the spread of infectious diseases. However, they also find that the negative effects of disease can be overcompensated by the benefits of being social. The authors stress that experimental studies are necessary to disentangle these effects. The study is of high standard and well-conducted. The take home message is clear and of general interest.

    The study highlights that experimental work is important to understand the relationship between parasitism, relatedness and living in groups. However, I missed an important aspect here. Experiments tend to stretch factors (sometimes to extremes), which may go square to the biology of the species. In some cases, this results in non-social organisms to be pressed in a group-environment. For example, the monoculture effect as we know it in agriculture is highly artificial. Clonal lines of crop are planted in high density, promising high yield, if pathogens stay out. These plants do not have a history of evolving mechanisms to deal with the effect of high relatedness. In contrast animals living in social groups, may never experience setting with non-relatives. Social insects evolved to deal with parasites by expressing specific adaptations, such a grooming, hygiene and social structure in the colony. Many social insects may never experience conditions of low relatedness. Thus, I expect it makes a difference if you experimentally force a non-social organism to be social, or a social organism to be asocial. I would be happy if this factor could be included in the reasoning, and maybe even analyzed quantitatively. For example, I would expect that non-social species made artificially to grow in groups of relatives, suffer much more from parasites than typical social animals with the same degree of relatedness.

    This is an important point. One of the main motivations for conducting this study was to test if species that typically live with kin have evolved adaptations to minimise any increase in susceptibility to pathogens brought about by living in groups with relatives. We therefore collected data on whether species are: a) typically social or non-social, and b) average levels of relatedness between individuals in groups under natural conditions (see Methods section ‘Data on species characteristics’).

    a) Testing differences between social and non-social species. All species included in our dataset had some part of their life-cycle where they were social (note we specifically excluded any studies on non-natural systems such as crops and domesticated species). This meant that only comparisons between species that are obligately social versus species that are social during specific life stages could made. This is problematic as assumptions need to be made about the strength of selection during different life cycle phases. For example, mortality caused by pathogens maybe particular high during the social juvenile phases of otherwise non-social species, resulting in selection for adaptions to reduce pathogen spread being similar to species that are obligately social. An additional problem was that experimental studies (a key factor highlighted by our analyses) of species that are non-social apart from specific life-cycle phases were rare (n=1, Rana latastei) precluding any meaningful comparisons.

    We have now added the following sentences to the methods to clarify this point:

    “We also collected data on whether species always lived in social groups (‘obligately social’) or whether species were only social during specific life stages (‘periodically social’). However, it was not possible to analyse this data as experimental manipulations of pathogens, a key factor influencing the relationship between relatedness and mortality and pathogen abundances, were only performed for one periodically social species (Rana latastei)” (Lines 425-430).

    b) Testing differences between species that typically live with kin and non-kin. The third aim of the paper was to test if species that typically live with kin have evolved to deal with pathogens as the referee suggests. We found that species that live with kin, such as social insects, have similar rates of mortality and pathogen abundances to species that live with non-kin (Figure 3). However, species that typically live with kin had lower rates of mortality in groups with higher relatedness when pathogens were absent compared to species that typically live with non-kin. This suggests that pathogens represent an omnipresent threat to all species, but that adaptations have evolved to reap the benefits of living with relatives in social species.

    In summary, as suggested by the referee we analysed whether “species made artificially to grow in groups of relatives, suffer much more from parasites than typical social animals with the same degree of relatedness” as much as was possible given the limitations of the published data. We have edited parts of the manuscript to emphasise that this was a key aim of the paper (Lines 66-74; 92-94; 136-153).

    The term (and concept) "monoculture" is typically used to describe clonal populations, predominantly in agricultural settings. I understand that the authors like to expand this term (as have others done before) to include social animals. However, for most people this would be a change in terminology and may cause misunderstandings. I would prefer if you could stick with the mainstream terminology and avoid pressing this concept into a new costume.

    We included the term “monoculture effect” to facilitate links to existing literature, both in the fields of agriculture and evolutionary biology (e.g. Ekroth at al 2019). While we think that making the reader aware of relevant work in other fields is valuable, we understand its prominence could give the impression that we included agricultural studies. Therefore, we have removed it from the abstract, but have chosen to keep one reference to the monoculture effect in the introduction.

    Reviewer #2:

    This study uses an unusually broad comparative data set to disentangle the positive (relatedness) and negative (pathogen pressure) effects of living in groups. The authors largely succeed in this task even though the data do not allow answers to all outstanding issues. Not unexpectedly, experimental manipulation studies appear to be most informative. The results are broadly consistent with expectations based on kin-selection theory and clarify the effects of a number of important covariables. The study is thoroughly executed and innovative in its approach. I expect this study to be interesting for a broad readership and this method of searching literature data to have considerable impact. Some suggestions strengthening this paper are below:

    • I think it would be helpful for readers to have the Discussion start with a few lines on what your study achieved in language that is complementary to the abstract, perhaps followed by a brief explanation of which angles/ambiguities/challenges you will be taking up in the paragraphs to follow.

    We have now edited the beginning of the discussion in accordance with this suggestion. It reads:

    “Our analyses show that pathogens can increase rates of mortality in groups of relatives. The detrimental effects of pathogens were, however, counteracted by high relatedness reducing mortality when pathogens were rare, particularly in species that live in kin groups. Such contrasting effects of relatedness meant that experimental manipulations were crucial for detecting the costs and benefits of living with relatives when the presence of pathogens varied. Additionally, high relatedness resulted in more even abundances of pathogens across groups, but more variable rates of mortality, highlighting the importance of population genetic structure in explaining the epidemiology of diseases. We discuss these findings in relation to the environments favouring the evolution of different social systems, the mechanisms that have evolved to prevent disease spread in social groups, and the types of study system where more experimental data are required” (Lines 171-181).

    • The rationale of this study is (often implicitly) that tendencies to live with relatives or not is a continuous variable. This surprised me because the senior author has written influential papers showing that family groups are different from non-family groups. In some contexts of this study it seems crucial to make that distinction. For example, a number of data points come from studies of social insects (bumblebees, honeybees, ants). Here, living with non-relatives is not an option but a given. It is well documented that these caste-differentiated colonies originated from ancestors that had exclusively full-sib colonies, so maximal relatedness was ancestral and became only diluted secondarily in some lineages. Would it be possible to check statistically whether the social insect data points always showed the same pattern as the other data points? That would test whether it matters that low relatedness is either derived or ancestral (as I think we implicitly assume to be the case in all other organisms).

    The primary studies included in our analyses were conducted on a diverse set of species where relatedness was often reported and measured on a continuous scale (range 0 to 1). Our rationale and statistical treatment of the data (the effect size of Pearson’s correlation coefficient captures continuous variation in relatedness) reflect the measures reported in the primary studies. This does not mean, however, that we believe groups evolve from along a continuum of within-group relatedness.

    As the referee points out there are two distinct routes to group formation that set the limits to relatedness within groups. In species, where offspring do not disperse from their natal patch (‘family’ groups) the opportunity for interacting with relatives is high, whereas in species where groups form after individuals disperse from natal patches (‘non-family’ groups) relatedness is typically low. Some variation in within-group relatedness subsequently arises within these two categories because of a number of modifying factors (breeder turnover, number of males and females founding groups, ‘budding’ dispersal and so on). However, the potential for kin selection to favour adaptations, including those that limit pathogen spread, remains fundamentally different between family and non-family groups. We tried to capture such differences by classifying species as typically living with kin and non-kin using life- history information (dispersal patterns, mating systems) and direct estimates of relatedness.

    We used the terms kin and non-kin rather than family and non-family because across such a diverse set of study species, with variable types of information (e.g. some species only had molecular genetic estimates of relatedness others had only life-history information), it was not possible to ascertain exactly how groups form for each species. Nevertheless, our analyses are aimed at addressing if species that typically live with kin, such as the social insects, have more effective mechanisms for reducing the impact of pathogens amongst relatives than species that live with non-kin.

    The referee makes an additional valuable point that for social insects ancestral levels of relatedness in groups are known to be high, with lower levels of relatedness being derived. Examining whether species with low versus high contemporary estimates of relatedness may therefore shed light on the importance of current versus past evolutionary responses to pathogens.

    Unfortunately, the sample sizes are just too limited to conduct any meaningful analyses. Only one species of social insect in our dataset was classified as living with non-kin (r <0.25). We also examined finer scale predictors of relatedness applicable to social insects (queen mating frequency: monogamous (r = 0.5) versus polyandrous (r > 0.25 & <0.5)). Sample sizes for crucial comparisons were again too small for formal analysis (Number of monogamous species with experimental data: pathogens present = 3, Pathogens absent = 3. Number of polyandrous species with experimental data: pathogens present = 2, Pathogens absent = 1).

    We have extended the discussion highlighting that more work on species with ancestral and derived levels of high and low levels of relatedness will aid our understanding of the evolutionary history of adaptations to minimise pathogen spread in groups (Lines 248-250). We have also checked and edited the manuscript to remove any implication that groups originate from a continuum of relatedness.

    • I wondered whether you could (interpretationally, i.e. in the discussion) do more with comparative data on pathogen pressure in the wild. The 1987 Hamilton chapter that you cite has lots of interesting natural history observations, which are now often supported by better data. I think he speculates about how altruistic soldiers evolved in aphids and thrips and connects their sociality with living in their own food (galls), which should mean low parasite pressure. The same is true for the lower termites. Would your results allow you to conjecture that all independent lineages that evolved differentiated castes (only possible in families with full siblings; or clones as in aphids) likely had to do that in disease free habitats?

    This is an interesting point and an area where further research would be very valuable. It fits in nicely with our current discussion of how the evolution of groups with high relatedness maybe more likely to occur in environments where pathogens are rare. This was rather vertebrate focused before and so we are grateful for the referee’s suggestion, which has broadened this point. The section now reads:

    “Parallel arguments have been made for social insects. Species with sterile worker castes, that only evolved in groups with high levels of relatedness, are thought to have arisen in environments protected from pathogens (Hamilton 1987). For example, sterile soldier castes have evolved at least six independent times in clonal groups of aphids, and the majority of these cases form galls that provide protection against pathogens (Hamilton, 1987; Stern and Foster, 1996). Escape from pathogens may therefore be a general feature governing the evolutionary origin, as well as the current ecological niches, of species living in highly related groups” (Lines 190-197).

    • I think some effort should be made to make Figures 2,3 and 4 easier to interpret. The ultra-brief acronyms along the y-axis take a while to digest and to realize the nestedness of the analyses. Could you give one piece of information on the left axis (spelled out like 'experimental data' and 'observational data' and the other piece on the right axis (spelled out as 'pathogens absent' and pathogens present'? It would also be helpful if the reader could fully understand the figures without first having to go through the entire method section, so I recommend you extend the legend to explain: 1. What Zr stands for. 2. What the directionality is (so the cryptic line just below Zr can become a proper sentence in the legend), and 3. The rationale of the multifactorial analyses with four or eight combinations (as you describe in the methods; I believe Figure 4 is an example of eight, but this remains rather hazy).

    Many thanks for these suggestions. We have now revised the axis labels and figure legends to improve interpretability.

  3. eLife

    Evaluation Summary:

    Group living may be beneficial for many reasons, but has costs in terms of increased rates of parasitism, in particular if group members are highly related. In this meta analysis, many original studies on questions related to parasitism, relatedness and group living are brought together in one unifying framework. The authors conclude that living in groups can indeed facilitate the spread of infectious diseases, but that these costs can be overcompensated by the benefits of group living.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #2 agreed to share their name with the authors.)

  4. eLife

    Reviewer #1 (Public Review):

    This meta analysis addresses a double-edged sword in evolutionary biology. Group living may be beneficial for many reasons, but has costs in terms of increased rates of parasitism. Furthermore, if groups are highly related, parasites that are genetically able to infect on member of the group may be able to infect all of them, putting the entire group at risk. In the her presented meta analysis, many original studies working on questions related to parasitism, relatedness and group living are brought together in one unifying framework. The authors find that indeed, group living can facilitate the spread of infectious diseases. However, they also find that the negative effects of disease can be overcompensated by the benefits of being social. The authors stress that experimental studies are necessary to disentangle these effects. The study is of high standard and well-conducted. The take home message is clear and of general interest.

    The study highlights that experimental work is important to understand the relationship between parasitism, relatedness and living in groups. However, I missed an important aspect here. Experiments tend to stretch factors (sometimes to extremes), which may go square to the biology of the species. In some cases, this results in non-social organisms to be pressed in a group-environment. For example, the monoculture effect as we know it in agriculture is highly artificial. Clonal lines of crop are planted in high density, promising high yield, if pathogens stay out. These plants do not have a history of evolving mechanisms to deal with the effect of high relatedness. In contrast animals living in social groups, may never experience setting with non-relatives. Social insects evolved to deal with parasites by expressing specific adaptations, such a grooming, hygiene and social structure in the colony. Many social insects may never experience conditions of low relatedness. Thus, I expect it makes a difference if you experimentally force a non-social organism to be social, or a social organism to be asocial. I would be happy if this factor could be included in the reasoning, and maybe even analyzed quantitatively. For example, I would expect that non-social species made artificially to grow in groups of relatives, suffer much more from parasites than typical social animals with the same degree of relatedness.

    The term (and concept) "monoculture" is typically used to describe clonal populations, predominantly in agricultural settings. I understand that the authors like to expand this term (as have others done before) to include social animals. However, for most people this would be a change in terminology and may cause misunderstandings. I would prefer if you could stick with the mainstream terminology and avoid pressing this concept into a new costume.

  5. eLife

    Reviewer #2 (Public Review):

    This study uses an unusually broad comparative data set to disentangle the positive (relatedness) and negative (pathogen pressure) effects of living in groups. The authors largely succeed in this task even though the data do not allow answers to all outstanding issues. Not unexpectedly, experimental manipulation studies appear to be most informative. The results are broadly consistent with expectations based on kin-selection theory and clarify the effects of a number of important covariables. The study is thoroughly executed and innovative in its approach. I expect this study to be interesting for a broad readership and this method of searching literature data to have considerable impact. Some suggestions strengthening this paper are below:

    - I think it would be helpful for readers to have the Discussion start with a few lines on what your study achieved in language that is complementary to the abstract, perhaps followed by a brief explanation of which angles/ambiguities/challenges you will be taking up in the paragraphs to follow.

    - The rationale of this study is (often implicitly) that tendencies to live with relatives or not is a continuous variable. This surprised me because the senior author has written influential papers showing that family groups are different from non-family groups. In some contexts of this study it seems crucial to make that distinction. For example, a number of data points come from studies of social insects (bumblebees, honeybees, ants). Here, living with non-relatives is not an option but a given. It is well documented that these caste-differentiated colonies originated from ancestors that had exclusively full-sib colonies, so maximal relatedness was ancestral and became only diluted secondarily in some lineages. Would it be possible to check statistically whether the social insect data points always showed the same pattern as the other data points? That would test whether it matters that low relatedness is either derived or ancestral (as I think we implicitly assume to be the case in all other organisms).

    - I wondered whether you could (interpretationally, i.e. in the discussion) do more with comparative data on pathogen pressure in the wild. The 1987 Hamilton chapter that you cite has lots of interesting natural history observations, which are now often supported by better data. I think he speculates about how altruistic soldiers evolved in aphids and thrips and connects their sociality with living in their own food (galls), which should mean low parasite pressure. The same is true for the lower termites. Would your results allow you to conjecture that all independent lineages that evolved differentiated castes (only possible in families with full siblings; or clones as in aphids) likely had to do that in disease free habitats?

    - I think some effort should be made to make Figures 2,3 and 4 easier to interpret. The ultra-brief acronyms along the y-axis take a while to digest and to realize the nestedness of the analyses. Could you give one piece of information on the left axis (spelled out like 'experimental data' and 'observational data' and the other piece on the right axis (spelled out as 'pathogens absent' and pathogens present'? It would also be helpful if the reader could fully understand the figures without first having to go through the entire method section, so I recommend you extend the legend to explain: 1. What Zr stands for. 2. What the directionality is (so the cryptic line just below Zr can become a proper sentence in the legend), and 3. The rationale of the multifactorial analyses with four or eight combinations (as you describe in the methods; I believe Figure 4 is an example of eight, but this remains rather hazy).

  6. Version published to 10.1101/2021.02.02.429439v1 on bioRxiv