DISCUSSION
In this study, we proposed a theoretical model to investigate how the intensity of host-switching, mediated by opportunity and compatibility of interaction (using host phylogenetic distance as proxy), shapes the phylogenetic history and ecology of the parasites. Our results indicated that different host-switching intensities can drive distinct ecological and evolutionary patterns of the parasite lineages. The model reproduced the eco-evolutionary pattern of all analysed empirical cases, but the range of predicted host-switching intensity varied among them. The parasitism type did not explain this variation, but the extent of the spatial scales of the empirical cases analysed present some correlation with the predicted intensity of host-switching. Host-switching intensity was higher for local empiric cases when compared to regional scale cases.
Variation in host-switching intensity through history influences the resulting eco-evolutionary patterns of the parasites involved. When the frequency of host-switching is sufficiently high, parasites can maintain the gene flow among infrapopulations of distinct hosts and speciation does not occur. On the other hand, if the gene flow is reduced, reduction in host-switching favours parasite speciation (see the eco-evolutionary dynamics in the movie available in supplemental information III S16 and S17). This was the general pattern observed for all simulated communities, strongly suggesting that host-switching is an important driver for parasite evolution. During simulations, in line with the oscillation hypothesis, species first increase their host repertoire (generalize) and then speciate (specialize) (Janz & Nylin 2008; Braga et al. 2018).
Our results also support that host phylogenetic relationship is a good predictor of host-switching - i. e. host-switching is most likely to occur between related host species. There is probably some common trait, such as a specific physiological mechanism that has evolved and may be being expressed through phylogeny. The dispersion of parasites followed by colonization of a new host lineage has been increasingly investigated in the Stockholm Paradigm (Agosta et al. 2010; Araujo et al. 2015; Nylin et al. 2018; Brooks et al. 2019). For a parasite lineage, the closer (phylogenetically) the species of the original and new host, the greater the possibility that the necessary combination of elements that compose the adequate resource is conserved or, is at least quali- and quantitatively similar. Hence, phylogenetic proximity has been widely recognized as a potential criterion to anticipate the emergence of new associations (Streicker et al. 2010; Damas et al. 2020). However, this is not a universal criterion, as compatibility appears to be also modulated by other biological elements (e.g. morphology, genetics, ecology) associated with the opportunity of encounters between hosts and parasites. This theoretical framework has provided evidence on how the host-switching mediated by compatibility and opportunity of interaction influence the dynamics of parasitic interactions leading to species diversification (Agosta et al. 2010; Nylin et al. 2018; Brooks et al. 2019). The possibility of encounter between potential symbionts in time and space, emerges from geographic distributions, ecology, and inherent biological traits of the associates. Compatibility emerges from the ancestral capacity in which both must be physiologically compatible to establish a long-term association. Compatibility and opportunity should occur simultaneously to allow the establishment of new associations. Thus, any factor that influences the compatibility and/or the opportunity among hosts and parasites may affect the intensity of host switching by parasites. Such factors can include biological and/or spatial variations, which may explain the varied predicted intensities of host-switching among analysed communities.
The parasitism type (ecto vs endoparasites) had no effect on the predicted intensity of host-switching by parasites. Ectoparasites have direct contact with the external environment, while endoparasites may have free-living infective stages but spend most of their life without direct contact with the environment (Bush et al. 2001). Because of this, higher host-switching intensity could be expected in ecto rather than endoparasites due to the amount of time under variable environmental conditions that could lead to a stronger selective pressure to use a broader array of hosts. In addition, all parasites analysed here are monoxenic and differ by the transmission strategy: ectoparasites can transmit between host individuals during adult stages, while the endoparasites only during the larval stage. Hence, the strategy of transmission of the ectoparasites allows successive host-switching by an individual parasite, while endoparasites are restricted to less host-switching events. This wider possibility of transmission was assumed to favor a higher intensity of host-switching by ectoparasites than by endoparasites (Boeger et al. 2005). Consequently, as we could not detect significant differences between the empirical networks analysed herein, our results do not support these hypotheses. For instance, an alternative explanation may be linked to the generalization that propagule size compensates for the wider possibility of transmission. Moreover, there is great heterogeneity in the characteristics of both ecto and endoparasites. Each parasitism type includes a great diversity of organisms, with profound differences in their evolutionary history and biological characteristics (for example, by comparing species of the genus fleas, lice, feather mites, helminths, platyhelminthes), which may be more influential to host-switching intensity than the general site of parasitism type itself. Expanding the analyses to a broader sample of empirical networks, including variations in the reproductive strategies may provide important insights on this question.
Unlike parasitism type, our results indicate that host-switching intensity is higher on local than regional spatial scale. The opportunity for interaction is increased in host communities at a local scale, as this reduces the likelihood that geographic barriers exist, hampering the possibility of encounter. This is evident when comparing rodent and flea cases at regional (ID. 4) and local spatial scales (ID. 5). Similarly, since the cases of Rhabdias spp. and frogs (ID. 8) are defined geographically (and not by host taxa) it was assumed that host-switching and ecological fitting were evolutionarily more important than association with particular host taxa (Kuzmin et al. 2014; Müller et al. 2018). However, the majority of the empirical cases analysed herein and elsewhere suggest that switches are responsible for parasite diversification and distribution (see Krasnov et al. 2016; Dona et al. 2017; Patella et al. 2017).
Although the model can reconstruct eco-evolutionary patterns of empirical cases, it has some limitations. First, there is no variation of the carrying capacity imposed by each host species; second, variation in abundance and spatial distribution are not explicitly considered; third, the parasites compete for the same resource (in contrast, each parasite species would have a carrying capacity); and finally, selective pressure is not explicitly modelled. The solutions to these limitations would make the model more realistic but also demand more computational time and increase the number of parameters to be evaluated. Thus, quantitative comparisons of predicted host-switching intensities should not be done, but the qualitative implications, as already discussed. Furthermore, we emphasize the need for empirical phylogenetic studies, since the availability of phylogenetic data on parasites is still scarce. Phylogenetic data on parasites are extremely important to clarify the role of host-switching in the ecological and evolutionary patterns of parasite lineages.
Our results indicate that the intensity of host-switching, mediated by opportunity and compatibility of interaction, influence the phylogenetic history and ecology of the parasites. As suggested by the Stockholm Paradigm, parasites may already carry the ancestral capabilities to switch to new hosts, and here we showed that the host evolutionary history, when associated with opportunity for contact and interaction compatibility (opportunity + compatibility are represented by the parameter r ), creates the possibility for the colonization of a new host species. Our study represents a first attempt to model and evaluate the empirical evolutionary history of hosts as a proxy for parasite resources and offers a new approach to understanding the eco-evolutionary patterns of parasite species. This model has important implications for predicting changes in host lineages in situations of environmental or climatic changes or yet in cases of emerging diseases where the parasitic host switches can cause disease outbreaks.