This paper examines patterns of cophylogenetic association between Drakaeninae orchids and mycorrhizal fungi (OMF). Since OMF can exist independently of the orchids, the authors hypothesize that duplications and host-switches are the main drivers accounting for phylogenetic congruence between both taxa.
This work is novel because of the system studied. The fact that OMF can optionally associated to orchids makes the system very appealing and its study can provide novel insight into coevolution of symbionts. Thus I believe that the paper can be of interest to a wide readership and inspire further work.
The paper sets clear hypotheses, is technically sound and the results are correctly interpreted in the light of current theory. I have no major objections to raise, but I think that it could benefit if the suggestions I provide below are considered.
My main concern stems from my utter ignorance of fungal isolation and culture, and might turn out as completely unfounded. However, since cophylogenetic analyses are sensitive to the quality of the host-symbiont association matrix, to which extent are the authors certain that there are no coverage errors (such as fungus species present but not detected or low phylogenetic resolution to separate sibling species) in their data? I think some critical assessment of data quality (concerning both fungal-orchid associations and phylogenetic trees) should be included in the paper. This is somehow addressed in the discussion (l. 409-413) but a more general remark in materials and methods would be useful to reassure the reader early about the correctness of the data used.
The paper reports new phylogenetic information of the Drakaeinae and their OMF, this should be better highlighted in the introduction. I suggest formulating this as a prior requirement to achieve the main objective (i.e. cophylogenetic analysis) of the paper.
The authors consider five macroevolutionary scenarios in host-symbiont systems (l. 57-65). There is actually a sixth often termed colonization (a.k.a host-sharing or incomplete host-switch). Colonization consists of a symbiont shifting from the original host to a new, unrelated one without the symbiont speciating in the new host. This event is quite common among pathogens that are able to infect phylogenetically distant hosts, and it is usually not well covered by event-based methods (Drinkwater et al. 2018 https://doi.org/10.48550/arXiv.1603.09415). For completeness, it should be included in the authors’ account.
Specific comments
Abstract l. 8: To be in line with l. 96-97 “congruence would arise from _duplications and_ host shifts”. [Note also that host shift/switch is used interchangeably in the manuscript. Since not everyone accepts that they are synonyms (Rózsa et al. 2015 https://doi.org/10.1017/CBO9781139794749.007), I would use host switch exclusively.]
l. 32 The abbreviations AM and EM are unnecessary.
l. 34 Not clear to me why the preceding phrase justifies a prediction of diffuse co-evolution.
l. 125-129 Reference to previous Durideae phylogenies does not seem to belong here. Perhaps move to materials and methods.
l. 448 “a high number of duplication events is unusual…” needs to be backed by reference(s)
l. 484-486 “While host-switching is common in many study systems, cospeciation is a rare event that can also generate congruency”. The use of “also” is awkward here. It seems to imply that host-switching leads to congruency as a general rule. However, host-switching is often invoked as a cause of incongruent cophylogenetic patterns. Only if it occurs among closely related host, the symbiont phylogeny would tend to mimic that of their hosts.
l. 538 “seems to be” seems to be better than “is” here.
Fig. 4. Since PACo yields a sum of squared residuals as statistic measuring overall phylogenetic congruence, the authors should consider reporting the square residuals instead of the plain residuals.
Fig. 3S was blank in my copy of the ms. and some other figures were clipped.