*4.2. Meta-Barcoding*

DNA was isolated using the DNeasy PowerFood Microbial DNA Isolation Kit (Qiagen Hilden, Germany) following the manufacturer's instructions. Bead-beating was carried out using a combination of 0.1 mm and 0.5 mm zirconia/silica beads (BioSpec Products, Butlersville, Oklahoma) in a Precellys Evolution homogenizer (Bertin Instruments, Montigny-le-Bretonneux, France) at 8000 RPM for 4 × 60 s. In order to prepare amplicons for sequencing, a two-step PCR was performed using sequences designed to amplify the fungal ITS region, while adding experiment-specific inline barcodes and appropriate adaptors for the Illumina sequencing platform. Briefly, first-round amplification of the ITS region was performed using the fungal-specific primers BITS (ACCTGCGGARGGATCA) and B58S3 (GAGATCCRTTGYTRAAAGTT) [18] which were modified to include both an inline barcode and Illumina adaptor sequences [19]. Second-round amplification added sequences required for Illumina dual-indexed sequencing via overhang PCR. Sequencing was performed using 2 × 300 bp chemistry (Ramaciotti Centre for Functional Genomics, Sydney, Australia). Paired-end reads were quality trimmed (Trimmomatic v0.38 [34]), adaptor trimmed (cutadapt v1.16 [35]) and merged into single synthetic reads (FLASH2 v2.2.00 [36]). Merged reads were de-replicated (USEARCH v10.0.240 [37]) and clustered (Swarm v2.2.2 [38]) into operational taxonomic units (OTUs) as presented previously [19]. Taxonomic annotation was performed against the UNITE database (qiime\_ver8\_dynamic\_02.02.2019) using a 98% similarity cut off (assign\_taxonomy.py module of QIIME v1.9.1 [39]). All sequence reads have been lodged in the NCBI database under the Bioproject accession PRJNA634973.
