2.2. Metagenomic DNA extraction, amplicon sequencing and bioinformatics 
Metagenomic DNA was extracted from 0.5 g of sample using Qiagen DNeasy PowerSoil Pro Kit (Carlsbad, CA, USA). The Internal Transcribed Sequence 1 ribosomal region (ITS1) and hypervariable region V4 of 16S ribosomal gene were targeted to assess the fungal and prokaryotic community composition, respectively. The ITS1 region was amplified using barcoded primers ITS1F/ITS2, suitable for shorter read length (Smith & Peay, 2014), while for the V4 region of 16S, barcoded F515/R806 primer set was used according to Caporaso et al. (2012). PCR reactions consisted of 1 μL of each primer, 12.5 μL of Taq DNA Polymerase (Thermo Fisher Scientific Inc., Waltham, MA, USA), 9.5 μL of nuclease-free water (Sigma–Aldrich, St. Louis, MO, USA) and 5 ng of DNA for a total volume of 25 μL and occurred in an automated thermal cycler (BioRad, Hercules, CA, USA),. The ITS1 locus and V4 region were amplified according to Coleine et al. (2021). Amplicons were quantified by a Qubit dsDNA HS Assay Kit (Life Technologies, Carlsbad, CA, USA) and then pooled. Paired-end sequencing (2 × 300 bp) was carried out on an Illumina MiSeq platform at the Edmund Mach Foundation (San Michele all’Adige TN, Italy). Demultiplexed ITS and 16S sequence datasets were processed using AMPtk (Palmer et al., 2018) v.1.5.1 software. Briefly, barcodes/indexes and primer sequences were removed from raw data. Reads were subjected to quality trimming to a maximum of 250 bp, discarding those shorter than 100 bp; sequencing artefacts were dropped by using USEARCH v.9.1.13 with default parameters (Edgar, 2010). Sequence quality filtering was performed with the expected error parameter of 0.9 (Edgar & Flyvbjerg, 2015); the cleaned reads were merged and clustered at 99% similarity using VSEARCH (Rognes et al., 2016) v.2.15.1, with DADA2 (Callahan et al., 2016). Global singletons and rare taxa (<5 reads) were skipped as likely false positives due to sequencing errors (Lindahl et al., 2013). Finally, taxonomic identification was performed with hybrid database SINTAX/UTAX (Edgar, 2010).