3.3 Gut fungal community variation in giant pandas after diet conversion
After quality filtering and assembly, 7,172,982 fungal ITS sequences were obtained from 108 fecal samples from the diet conversion experiments. The sequences were clustered into 15,547 OTUs at the 97% sequence identity threshold. Although richness and diversity varied with diet conversions, significant differences in fungal community richness and diversity were not observed among communities from the three experimental groups (p > 0.05, ANOVA) (Figure 1a, b). However, PCoA indicated that the communities from hosts with the same diet clustered together and separately from others (Figure 1c). Likewise, significant differences in community structure based on BC distances were identified among the three groups (p < 0.05, AMOVA).
Ascomycota was the most dominant phyla among communities in the diet conversion experiment, followed by Basidiomycota (Figure 1d). Basidiomycota notably increased markedly in the OBD (41.9%) communities relative to the OMD (7.6%) communities (Non-parametric factorial Kruskal-Wallis sum-rank test, LDA>4) (Figure 1e). The 10 most abundant fungal genera in each group were further investigated, excluding unidentified genera. Candida (37.0%),Saccharomyces (6.2%), and Microidium (6.1%) were the three most abundant genera among communities of the OMD treatment.Candida (3.0%), Microidium (2.7%), andGibberella (1.0%) were the three most abundant genera in the MBD communities. Lastly, Cystofilobasidium (9.0%), Guehomyces(8.1%), and Microidium (5.2%) were the three most abundant genera in the OBD communities (Figure 1f). Candida abundances were significantly higher in the OMD (37.0%) communities and were significantly lower in the OMD (37.0%) to OBD (0.1%) communities (Non-parametric factorial Kruskal-Wallis sum-rank test, LDA>4) (Figure 1e, f). In addition,Cystofilobasidium , Guehomyces, and Gibberella were significantly more abundant in the OBD communities (9.0%, 8.1%, and 2.8%, respectively) than in the OMD communities (0.03%, 0.0%, and 0.01%, respectively) (Non-parametric factorial Kruskal-Wallis sum-rank test, LDA>4).