DNA extraction, PCR amplification and sequencing
We extracted DNA from each bulk subsample using the Biosprint 96 DNA Blood Kit (Qiagen) on a Thermo KingFisher Flex automated extraction instrument. We quantified the DNA concentration using a NanoDrop spectrophotometer and combined the extracts of each pair of subsamples at a 1:10 ratio (Arribas et al., 2021b). Then, we PCR amplified the 3’ end of the cytochrome c oxidase subunit I (COI) barcode region corresponding to the 418 bp bc3’ fragment using the Ill_B_F (Shokralla et al., 2015) and Fol-degen-rev (Yu et al., 2012) primers. Amplifications were performed following the PCR conditions described in Arribas et al. (2016). We carried out five independent PCR replicates per sample, visualised the PCR products on an agarose gel and pooled the three amplicons per sample that presented the brightest bands, which were purified using a magnetic bead-based protocol (Agencourt AMPure XP). We included three negative controls corresponding to different wet-lab steps (lysis, DNA extraction and PCR amplification). The 88 metabarcoding samples and the three negative controls were used for a dual-indexed library preparation following the Nextera XT DNA workflow (llumina, San Diego, CA, USA) and were sequenced on a paired-end 2x300 bp lane of an Illumina MiSeq platform at the Earlham Institute (Norwich, UK).
We individually extracted DNA from each of the ‘voucher’ specimens as described above. We amplified the Folmers COI barcode region (658 bp, overlapping with the 418 bp metabarcoding fragment), using the primers Fol‐degen‐for and Fol‐degen‐rev (Yu et al., 2012) and following the PCR conditions described in Arribas et al. (2016). The PCR products were purified and bidirectionally Sanger sequenced (Macrogen, South Korea). The generation of a curated ‘voucher’ reference catalogue allowed us to improve the taxonomic assignment of the metabarcode reads and to apply a recently developed method for read filtering (metamate, Andújar et al., 2021, see below).