DNA extraction and genotyping

DNA was individually extracted using the QIAamp DNA Mini Kit (QIAGEN) following manufacturer’s instructions. DNA integrity was checked by gel electrophoresis and quantified by NanoDrop or Qubit. To build individual genomic libraries with 2b-RAD we followed the protocol described in Barbanti and co-authors¹⁶. In brief, the Genomic DNA of each sample was individually digested by AlfI. Adaptors with degenerate bases (5′-NN-3′) were ligated to the sticky ends of the digested sequences. Barcodes and Illumina primers were attached to the adaptors by PCR. Purification was performed using magnetic beads, quantified using PicoGreen® and samples pooled and single end sequenced (50 bp) with a HiSeq 2500 Illumina at the Center for Genomic Regulations (CRG) of Barcelona. Some of the settlers of Blanes (N = 12) and Xabia (N = 12) were sequenced previously using the same methodology¹⁶ and data combined with new data from the present study for additional 37 settlers and 44 survivors. Several previously sequenced samples were repeated and showed high genotyping accuracy (mean of 98.6% coincident genotypes).

Raw sequences of all individuals used in the analysis were processed simultaneously prior to genotyping, using customized scripts¹⁶. Those files were used for genotyping with STACKS vs 2.53 software⁴⁹. To construct the catalog of loci (i.e. the list of all loci across individuals as defined by the program) two mismatches were allowed between stacks within (M = 2) and between (n = 2) individuals, and a minimal stack depth of three was required (m = 3). Individual genotypes were outputted as SNP VCF files. Additional filters were applied using VCFtools vs 1.12⁵⁰. Individual genotypes with a depth below 5X were not considered. Loci with a missingness value higher than 30%, and loci with a minimum allele frequency (MAF) of 0.05 were removed from the dataset with VCFtools. We also removed loci with high mean depth across all individuals, corresponding to a sequencing depth above 1.5 times the interquartile range from the dataset. Finally, the loci in HW disequilibrium at two of the three localities were discarded to remove possible paralogous loci.