136
younger, the vertebrate stem lineage extends well back into the early Cambrian or
even Ediacaran. The tunicate-vertebrate divergence, for instance, is resolved at 575
and 550 Ma (by Dohrmann and Wörheide 2017 and dos Reis et al. 2015, respectively), close to the Cryogenian estimate for the inferred endocnidozoan divergence.
This provides evidence that fish ancestors were likely present and could equally
have been first hosts. The scenario that very ancient endocnidozoans may have been
free-living is also not recognised. So far there is no fossil evidence to demonstrate
when and in what hosts endocnidozoans or myxozoans first appeared. Further processes that are not considered include host switching and extinction. Myxozoans
clearly do undergo host switching and presently known hosts may have been
acquired at any time since they have evolved. However, under Holzer et al.’s scenario, myxozoans have consistently only exploited those host lineages in which they
are currently parasites.
Cophylogenetic studies conducted by Holzer et al. (2018) are employed along
with their molecular clock analyses to argue that invertebrates were adopted as first
hosts. However, the inferences proposed as a result of these analyses are potentially
problematic. Thus, Holzer et al. (2018) infer that invertebrate host and parasite phylogenies are congruent ‘down to the most basal branches’ and highlight that all
major myxozoan clades are associated with different invertebrate acquisition events
(phylactolaemates, oligochaetes and polychaetes) (apart from Sphaerospora whose
invertebrate hosts are unknown). Yet all major clades (including Sphaerospora) are
also associated with fish hosts and, as argued above, fish (or their ancestors) cannot
necessarily be excluded as first hosts. Furthermore, relationships among phylactolaemate bryozoans are largely unresolved in molecular phylogenetic analyses
(most genera and families in this small group comprise a polytomy; Waeschenbach
et al. 2012), and recent studies reveal new higher level taxa in undersampled environments (e.g. Amazonia; Wood and Okamura 2017). Thus the congruence of phylogenies may be anomalous and compromised by both limited sampling and poor
phylogenetic resolution of invertebrate hosts. Finally, their multiple examples of
congruent cophylogenies of myxozoans and fish hosts could arise if fish served as
primary hosts followed by host switching, rather than being multiply acquired as
secondary hosts as proposed. Quite apart from these specific issues are more general
concerns about what can be inferred by such cophylogenetic investigations. These
include the problem of non-independence of phylogenies using PARAFIT as done
by Holzer et al. (2018) (Felsenstein 1985; de Vienne et al. 2013), testing for congruence on the basis of estimated phylogenies without accounting for uncertainty in the
inference (Warnock and Engelstädter 2021) and underestimation of the high probability of host-shift speciations (de Vienne et al. 2013).
4.6.2.3 The Endocnidozoa
Despite their molecular clock evidence for a common origin of Myxozoa/Polypodium
in the late Cryogenian, Holzer et al. (2018) propose that Polypodium and myxozoans are actually not sister taxa and that Endocnidozoa is thus an invalid taxon. They
B. Okamura and A. Gruhl
younger, the vertebrate stem lineage extends well back into the early Cambrian or
even Ediacaran. The tunicate-vertebrate divergence, for instance, is resolved at 575
and 550 Ma (by Dohrmann and Wörheide 2017 and dos Reis et al. 2015, respectively), close to the Cryogenian estimate for the inferred endocnidozoan divergence.
This provides evidence that fish ancestors were likely present and could equally
have been first hosts. The scenario that very ancient endocnidozoans may have been
free-living is also not recognised. So far there is no fossil evidence to demonstrate
when and in what hosts endocnidozoans or myxozoans first appeared. Further processes that are not considered include host switching and extinction. Myxozoans
clearly do undergo host switching and presently known hosts may have been
acquired at any time since they have evolved. However, under Holzer et al.’s scenario, myxozoans have consistently only exploited those host lineages in which they
are currently parasites.
Cophylogenetic studies conducted by Holzer et al. (2018) are employed along
with their molecular clock analyses to argue that invertebrates were adopted as first
hosts. However, the inferences proposed as a result of these analyses are potentially
problematic. Thus, Holzer et al. (2018) infer that invertebrate host and parasite phylogenies are congruent ‘down to the most basal branches’ and highlight that all
major myxozoan clades are associated with different invertebrate acquisition events
(phylactolaemates, oligochaetes and polychaetes) (apart from Sphaerospora whose
invertebrate hosts are unknown). Yet all major clades (including Sphaerospora) are
also associated with fish hosts and, as argued above, fish (or their ancestors) cannot
necessarily be excluded as first hosts. Furthermore, relationships among phylactolaemate bryozoans are largely unresolved in molecular phylogenetic analyses
(most genera and families in this small group comprise a polytomy; Waeschenbach
et al. 2012), and recent studies reveal new higher level taxa in undersampled environments (e.g. Amazonia; Wood and Okamura 2017). Thus the congruence of phylogenies may be anomalous and compromised by both limited sampling and poor
phylogenetic resolution of invertebrate hosts. Finally, their multiple examples of
congruent cophylogenies of myxozoans and fish hosts could arise if fish served as
primary hosts followed by host switching, rather than being multiply acquired as
secondary hosts as proposed. Quite apart from these specific issues are more general
concerns about what can be inferred by such cophylogenetic investigations. These
include the problem of non-independence of phylogenies using PARAFIT as done
by Holzer et al. (2018) (Felsenstein 1985; de Vienne et al. 2013), testing for congruence on the basis of estimated phylogenies without accounting for uncertainty in the
inference (Warnock and Engelstädter 2021) and underestimation of the high probability of host-shift speciations (de Vienne et al. 2013).
4.6.2.3 The Endocnidozoa
Despite their molecular clock evidence for a common origin of Myxozoa/Polypodium
in the late Cryogenian, Holzer et al. (2018) propose that Polypodium and myxozoans are actually not sister taxa and that Endocnidozoa is thus an invalid taxon. They
B. Okamura and A. Gruhl