Office Tool V5.9.3.3 (Installation Activation)
([@B44]). All metagenomic datasets used in this study are available at the EBI metagenomics portal ([@B4]).
Bioinformatic analysis and taxonomic and functional annotation of metagenomes. {#s4.3}
------------------------------------------------------------------------------
Metagenomic data were analyzed as described previously ([@B38]). Briefly, following quality control of raw reads and dereplication ([@B45]), we performed taxonomic annotation for genomes against the IMG/M database using the "taxo2ref" function, followed by dereplication using the "ref2c" function. For phylogenetic analysis, we generated a maximum-likelihood phylogenetic tree using RAxML v8 ([@B46]) with the PROTCATWAG model and a bootstrap of 100. Functional annotation was performed against the Kyoto Encyclopedia of Genes and Genomes (KEGG) ([@B47]) with the "hmmscan" function, using the model mmseqid16. The KEGG modules were used to perform KEGG pathway annotation. Annotation of protein-coding genes was performed using the RAST ([@B48]) server.
Microbial selection at 16S rRNA and functional gene level. {#s4.4}
----------------------------------------------------------
We selected OTUs shared between microbial communities of freshwater and marine source water to find the taxonomic group that drives the main variations in the microbial communities. We first calculated pairwise Spearman correlations between different environmental and microbial factors using the "corrplot" package in R. We then calculated the percentage of each microbial factor (OTU, taxonomic rank, and functional category) for each source of water. We performed OTU or taxonomic rank-based multivariate analysis of variance (MANOVA) to select the microbial factors that drive the difference in microbial community structure. We used "ade4" package in R to analyze the Bray-Curtis dissimilarity among source-to-transmission samples. We performed singular value decomposition (SVD) on the Bray-Curtis dissimilarity matrix and selected the first three components that best represent the data. We used the permutation-based ANOVA (PERMANOVA) to test the significance of pairwise differences among source-to-transmission communities. We also performed the analysis at higher taxonomic ranks (i.e., class, order, family, and genus) and functional groups (i.
44926395d7

Related