To characterize microbial communities and to perform functional analyses, the following wrappers have been employed: assign_taxonomy (to assign Greengenes taxonomy to OTUs), make_otu_table (to create biom file with OTUs and taxonomy), compute_core_microbiome (to calculate the core microbiome), summarize_taxa_through_plots (to produce the taxonomical files and charts), alpha_rarefaction and beta_diversity_through_plots (to assess respectively the alpha- and beta-rarefaction diversity indices), principal_coordinates.py (to compare groups of samples based on phylogenetic distance metrics). Scatterplot3d package for R was employed for the visualization of PCoA data. To compare sampling seasons (summer versus winter) and sampling sites (PM versus “other sites”) within each sampling season, ANOVA analyses have been performed at genus level (P-value < 0.005; FDR < 0.01) using CALYPSO version 7.18. PICRUSt (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States) was used for inference of microbial metabolic pathways starting from the taxonomic composition estimated from 16S rDNA data (Langille et al., 2013). Finally, STAMP (Statistical Analysis of Metagenomic Profiles) software was used for statistical analyses performing the Whelch's t-test (Parks et al., 2014) on the PICRUSt data.