2.2. Sequencing
A restriction site-associated DNA library was generated by GBS method (Elshire et al., 2011). To select for the best enzyme with reproducible genomic fragments across samples (a critical criterion for GBS experiments), we first generated a pilot library (four samples) with Illumina MiSeq (150 bp), using two common restriction enzymes, Pstl and ApekI. PstI provided the best reproducibility and number of variants and was therefore chosen to generate the final library.
Genomic DNA was digested with the PstI, fragments were tagged with individual barcodes, PCR-amplified, multiplexed, and sequenced on dual lanes on an Illumina HiSeq 2500 platform (2x100 bp). Raw read sequences were demultiplexed and quality checked using FastQC v0.10.1 (Wingett and Andrews, 2018); available athttps://www.bioinformatics.babraham.ac.uk/proj‐ects/fastqc/).
Additionally, raw RADseq genomic data of P. lilfordi generated on a previous study (Bassitta et al., 2021) were downloaded from the online repository (PRJNA645796, 91 samples from 10 populations). The GBS dataset (this study) and the RADseq dataset (Bassitta et al., 2021) shared the restriction enzyme PstI, which allowed data integration for comparative purposes (see below). Two localities of Colom and Aire were sampled and sequenced in both studies (different individuals) (Table 1 and S1).