The interplay of fungal and bacterial microbiomes on rainforest frogs following a disease outbreak

Abstract

Emerging infectious diseases are a serious threat to wildlife populations, and there is growing evidence that host microbiomes play important roles in infection dynamics, possibly even mitigating diseases. Nevertheless, most research on this topic has focused only on bacterial microbiomes, while fungal microbiomes have been largely neglected. To help fill this gap in our knowledge, we examined both the bacterial and fungal microbiomes of four sympatric Australian frog species, which had different population‐level responses to the emergence of chytridiomycosis, a widespread disease caused by the fungal pathogen Batrachochytrium dendrobatidis (Bd). We sequenced 16,884 fungal amplicon sequence variants (ASVs) and 41,774 bacterial ASVs. Bacterial communities had higher richness and were less variable within frog species than were fungal communities. Nevertheless, both communities were correlated for both ASV richness and beta diversity (i.e., frogs with similar bacterial richness and community composition tended to also have similar fungal richness and community composition). This suggests that either one microbial community was having a large impact on the other or that they were both being driven by similar environmental factors. For both microbial taxa, we found little evidence of associations between Bd (prevalence or intensity) and either individuals' ASVs or beta diversity. However, there was mixed evidence of associations between richness (both bacterial and fungal) and Bd, with high richness potentially providing a protective effect. Surprisingly, the relative abundance of bacteria that have previously been shown to inhibit Bd was also positively associated with Bd infection intensity, suggesting that a high relative abundance of those bacteria provides poor protection against infection.

The skin microbiota of the axolotl Ambystoma altamirani is highly influenced by metamorphosis and seasonality but not by pathogen infection

“…It has been shown that Bd presence have contrasting effects over skin microbiota diversity inducing changes in skin microbiota composition following infection [ 21 , 23 , 71 ] or not influencing diversity of skin microbial communities [ 42 , 44 ] as we found in this study. However, it also has been shown that relative abundances of some bacterial members of the skin microbiota correlates with chytrid infection intensity [ 19 , 44 , 45 ] and its suggested that according to the type of correlation these groups could act as anti Bd bacteria [ 19 ].…”

“…Considering that the main source of diversity of the skin microbiota are the environmental microbial communities and that they vary in response to environmental variation [ 65 – 67 ] it is likely that the skin microbiota reflect to some extent the environmental variations across localities as seen in the case of pre-metamorphic axolotls. It has been shown that skin bacterial diversity vary in response to precipitation [ 19 , 23 ] temperature [ 22 ] or elevation gradients [ 24 , 41 , 42 ]. However, genetic differences across populations could also explain some of our results, since a previous study showed that A. altamirani populations of sites 2 and 3 are genetically differentiated [ 68 ].…”

“…In temperate regions, where the four seasons are well defined through the year, seasonal changes have been linked to the temporal dynamics of the amphibian skin microbiota [ 22 , 37 – 39 ]. In addition, it has been shown that spatial variation such as elevation gradients [ 40 – 42 ] or distinct microhabitats [ 43 ] influence the skin microbial diversity of amphibians.…”

Optimising High-throughput sequencing data analysis, from gene database selection to the analysis of compositional data: A case study on tropical soil nematodes

“…On the other hand, the clr-transformation based methods might overestimate tested effects: these methods do not account for uneven library size across samples (Kaul et al, 2017), potentially magnifying differences among samples. Log-transformation based methods likely to exaggerate the importance of rare taxa, while rarefaction or proportion based methods, which fully normalize read depths across samples, may be superior to log-transformation based methods for comparing ecological communities (McKnight et al 2019). Although developed for differential abundance testing, ANCOM-BC methods outperformed other normalization methods, including the proportion-based method proposed by McKnight et al (2019), for the visualization of separate samples from different populations (Lin & Peddada, 2020).…”

“…Log-transformation based methods likely to exaggerate the importance of rare taxa, while rarefaction or proportion based methods, which fully normalize read depths across samples, may be superior to log-transformation based methods for comparing ecological communities (McKnight et al 2019). Although developed for differential abundance testing, ANCOM-BC methods outperformed other normalization methods, including the proportion-based method proposed by McKnight et al (2019), for the visualization of separate samples from different populations (Lin & Peddada, 2020). It is beyond the scope of the present study to statistically demonstrate the superiority of ANCOM-BC based methods, but our results do highlight the impacts of analyzing marker-gene data using different data normalization methods as well as the potential weaknesses of rarefaction-and log-transformation based methods for analyzing nematode HTS data.…”