Limits...
Specialization for resistance in wild host-pathogen interaction networks.

Barrett LG, Encinas-Viso F, Burdon JJ, Thrall PH - Front Plant Sci (2015)

Bottom Line: At the individual level, specialization was strongly linked to partial resistance, such that partial resistance was effective against a greater number of pathogens compared to full resistance.Second, we found that all networks were significantly nested.Third, we found that resistance networks were significantly modular in two spatial networks, clearly reflecting spatial and ecological structure within one of the networks.

View Article: PubMed Central - PubMed

Affiliation: Commonwealth Scientific and Industrial Research Organization Agriculture Flagship Canberra, ACT, Australia.

ABSTRACT
Properties encompassed by host-pathogen interaction networks have potential to give valuable insight into the evolution of specialization and coevolutionary dynamics in host-pathogen interactions. However, network approaches have been rarely utilized in previous studies of host and pathogen phenotypic variation. Here we applied quantitative analyses to eight networks derived from spatially and temporally segregated host (Linum marginale) and pathogen (Melampsora lini) populations. First, we found that resistance strategies are highly variable within and among networks, corresponding to a spectrum of specialist and generalist resistance types being maintained within all networks. At the individual level, specialization was strongly linked to partial resistance, such that partial resistance was effective against a greater number of pathogens compared to full resistance. Second, we found that all networks were significantly nested. There was little support for the hypothesis that temporal evolutionary dynamics may lead to the development of nestedness in host-pathogen infection networks. Rather, the common patterns observed in terms of nestedness suggests a universal driver (or multiple drivers) that may be independent of spatial and temporal structure. Third, we found that resistance networks were significantly modular in two spatial networks, clearly reflecting spatial and ecological structure within one of the networks. We conclude that (1) overall patterns of specialization in the networks we studied mirror evolutionary trade-offs with the strength of resistance; (2) that specific network architecture can emerge under different evolutionary scenarios; and (3) network approaches offer great utility as a tool for probing the evolutionary and ecological genetics of host-pathogen interactions.

No MeSH data available.


Related in: MedlinePlus

Bipartite networks for the Linum-Melampsora temporal datasets. Black squares indicate full resistance, gray squares indicate partial resistance, and white squares indicate full susceptibility. Each pair of panels represents a separate population/temporal network (labeled Kiandra, N1, etc.). The two columns for each population pair are formed from the same base dataset but sorted in different ways so as to facilitate visually testing different evolutionary hypotheses. The left panels have been sorted according to sampling period (i.e., temporally), while the panels to the right have been sorted to maximize nestedness. For the temporally sorted networks, the red line demarks contemporary host and pathogens, such that an interactions above and to the left of the red line indicate hosts interacting with past pathogens (year of sampling is shown in the bottom left panel). The expectation under an arms race hypothesis is that there should be a nested pattern emerging over time, such that more resistance reactions occur above and to the left of the red line. All networks were significantly nested when sorted to maximize nestedness (Table 3: NODF; P < 0.05), but none were significantly nested under the temporal arrangement.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4585140&req=5

Figure 5: Bipartite networks for the Linum-Melampsora temporal datasets. Black squares indicate full resistance, gray squares indicate partial resistance, and white squares indicate full susceptibility. Each pair of panels represents a separate population/temporal network (labeled Kiandra, N1, etc.). The two columns for each population pair are formed from the same base dataset but sorted in different ways so as to facilitate visually testing different evolutionary hypotheses. The left panels have been sorted according to sampling period (i.e., temporally), while the panels to the right have been sorted to maximize nestedness. For the temporally sorted networks, the red line demarks contemporary host and pathogens, such that an interactions above and to the left of the red line indicate hosts interacting with past pathogens (year of sampling is shown in the bottom left panel). The expectation under an arms race hypothesis is that there should be a nested pattern emerging over time, such that more resistance reactions occur above and to the left of the red line. All networks were significantly nested when sorted to maximize nestedness (Table 3: NODF; P < 0.05), but none were significantly nested under the temporal arrangement.

Mentions: When matrices were sorted to maximize nestedness, values of NODF were significantly higher than expected by chance for all networks (Table 3) and all networks visually appeared to be nested (Figures 3–5). This indicates that for all 8 networks, resistance specificities displayed by hosts with a wide range of resistance (generalists) tended to encompass resistance specificities displayed by hosts with a narrow range of resistance (specialists). In addition, we note that in networks with full resistance types, generalist hosts tended to carry relatively high levels of full resistance (Figures 3, 4; although this pattern is less evident in the BH network).


Specialization for resistance in wild host-pathogen interaction networks.

Barrett LG, Encinas-Viso F, Burdon JJ, Thrall PH - Front Plant Sci (2015)

Bipartite networks for the Linum-Melampsora temporal datasets. Black squares indicate full resistance, gray squares indicate partial resistance, and white squares indicate full susceptibility. Each pair of panels represents a separate population/temporal network (labeled Kiandra, N1, etc.). The two columns for each population pair are formed from the same base dataset but sorted in different ways so as to facilitate visually testing different evolutionary hypotheses. The left panels have been sorted according to sampling period (i.e., temporally), while the panels to the right have been sorted to maximize nestedness. For the temporally sorted networks, the red line demarks contemporary host and pathogens, such that an interactions above and to the left of the red line indicate hosts interacting with past pathogens (year of sampling is shown in the bottom left panel). The expectation under an arms race hypothesis is that there should be a nested pattern emerging over time, such that more resistance reactions occur above and to the left of the red line. All networks were significantly nested when sorted to maximize nestedness (Table 3: NODF; P < 0.05), but none were significantly nested under the temporal arrangement.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4585140&req=5

Figure 5: Bipartite networks for the Linum-Melampsora temporal datasets. Black squares indicate full resistance, gray squares indicate partial resistance, and white squares indicate full susceptibility. Each pair of panels represents a separate population/temporal network (labeled Kiandra, N1, etc.). The two columns for each population pair are formed from the same base dataset but sorted in different ways so as to facilitate visually testing different evolutionary hypotheses. The left panels have been sorted according to sampling period (i.e., temporally), while the panels to the right have been sorted to maximize nestedness. For the temporally sorted networks, the red line demarks contemporary host and pathogens, such that an interactions above and to the left of the red line indicate hosts interacting with past pathogens (year of sampling is shown in the bottom left panel). The expectation under an arms race hypothesis is that there should be a nested pattern emerging over time, such that more resistance reactions occur above and to the left of the red line. All networks were significantly nested when sorted to maximize nestedness (Table 3: NODF; P < 0.05), but none were significantly nested under the temporal arrangement.
Mentions: When matrices were sorted to maximize nestedness, values of NODF were significantly higher than expected by chance for all networks (Table 3) and all networks visually appeared to be nested (Figures 3–5). This indicates that for all 8 networks, resistance specificities displayed by hosts with a wide range of resistance (generalists) tended to encompass resistance specificities displayed by hosts with a narrow range of resistance (specialists). In addition, we note that in networks with full resistance types, generalist hosts tended to carry relatively high levels of full resistance (Figures 3, 4; although this pattern is less evident in the BH network).

Bottom Line: At the individual level, specialization was strongly linked to partial resistance, such that partial resistance was effective against a greater number of pathogens compared to full resistance.Second, we found that all networks were significantly nested.Third, we found that resistance networks were significantly modular in two spatial networks, clearly reflecting spatial and ecological structure within one of the networks.

View Article: PubMed Central - PubMed

Affiliation: Commonwealth Scientific and Industrial Research Organization Agriculture Flagship Canberra, ACT, Australia.

ABSTRACT
Properties encompassed by host-pathogen interaction networks have potential to give valuable insight into the evolution of specialization and coevolutionary dynamics in host-pathogen interactions. However, network approaches have been rarely utilized in previous studies of host and pathogen phenotypic variation. Here we applied quantitative analyses to eight networks derived from spatially and temporally segregated host (Linum marginale) and pathogen (Melampsora lini) populations. First, we found that resistance strategies are highly variable within and among networks, corresponding to a spectrum of specialist and generalist resistance types being maintained within all networks. At the individual level, specialization was strongly linked to partial resistance, such that partial resistance was effective against a greater number of pathogens compared to full resistance. Second, we found that all networks were significantly nested. There was little support for the hypothesis that temporal evolutionary dynamics may lead to the development of nestedness in host-pathogen infection networks. Rather, the common patterns observed in terms of nestedness suggests a universal driver (or multiple drivers) that may be independent of spatial and temporal structure. Third, we found that resistance networks were significantly modular in two spatial networks, clearly reflecting spatial and ecological structure within one of the networks. We conclude that (1) overall patterns of specialization in the networks we studied mirror evolutionary trade-offs with the strength of resistance; (2) that specific network architecture can emerge under different evolutionary scenarios; and (3) network approaches offer great utility as a tool for probing the evolutionary and ecological genetics of host-pathogen interactions.

No MeSH data available.


Related in: MedlinePlus