Statistical analyses
Statistical analyses were performed based on the results of the OTU clustering and species composition analyses. Sequences with an OTU threshold < 10 were discarded. OTUs with an identity value of ≥ 99% and an E-value of ≥ 10−5 were screened and the OTUs from the same species were combined. If an OTU could not be compared at the species level, statistics were created at the next level, such as genus or family. Microsoft Excel was used to determine the proportion of valid sequences for each fish sample.
The Shannon-Wiener diversity index (H), Pielou uniformity index (J), richness index, and Simpson Dominance Index (D) were used to assess community diversity. To explore the differences or similarities in community composition between different groups of samples, the number of OTUs was used to calculate the Bray-Curtis distance measure matrix for nonmetric multidimensional scaling (NMDS), permutational multivariate analysis of variance (PerMANOVA), and analysis of similarities (ANOSIM). Additionally, the relationship between environmental factors and community composition was investigated using canonical correspondence analysis (CCA). Pearson correlation analyses were used to investigate the influence of environmental and human factors on the species richness index.
All analyses were conducted using R v2.5.6, with the vegan package used for NMDS, PerMANOVA, ANOSIM, and diversity index calculations. The Pacman package was used for the Pearson correlation analysis of environmental factors, whereas diversity index boxplots were generated using the Amplicon package. ANOVA was performed using Perim (v9.1.0) software. The GAMs were implemented using the function “gam” with the default settings in the R package “mgcv.”