Soil biota sampling
Soil biodiversity was measured using environmental DNA (eDNA) metabarcoding from the soil samples. In the field, we extracted eDNA from a 15 g aliquot following Taberlet et al. (2012, 2018). The remaining soil was sieved at 2 mm and used to measure soil physico-chemical properties (Table 1).
We employed six DNA markers, including two universal markers (euka02 for eukaryote, bact01 for bacteria) and four clade-specific markers (fung02 for fungi, inse01 for insect, olig01 for oligochaeta, and coll02 for collembola) (Taberlet et al. 2018). Details regarding the markers and molecular analyses are detailed in Appendix 2. A standardised bioinformatic pipeline was applied to curate the retrieved sequences from contaminants and errors (Calderón‐Sanou et al. 2020), using the OBITools software (Boyer et al. 2016) and the ‘metabaR’ R package (R Core Team 2020; Zinger et al. 2021). Sequences were clustered into Molecular Operational Taxonomic Unit (MOTU) with SUMACLUST (Mercier et al. 2013) using a clustering threshold of 97% for Euka02, Fung01 and Bacte01, 85% for Coll01, 88% for Olig01 and 95% for Inse01, following the recommendations in Bonin et al. (2023). Taxonomic annotations were performed on MOTUs with the SILVAngs pipeline (Quast et al. 2013) (SILVA version 138.1) for ribosomal universal markers. For clade specific markers, we used the ecotag program from the OBITools, and marker-specific databases built with the ecoPCR program (Ficetola et al. 2010), from the EMBL database version 136. Metazoan taxa not registered in the European region were removed from the data (<1% of total reads) using the GBIFfilter tool (https://github.com/nleguillarme/gbif-filter-python).