2.3 | Genotyping-by-Sequencing Library and SNPs Selection
We collected leaf tissue samples from 140 individuals of J. chilensis (juveniles and adults) spaced within 150-500 m, between February and December 2015. Samples were stored in silica. For molecular analysis, DNA was extracted using the DNeasy Plant Mini Kit (Qiagen, EUA). DNA concentration was quantified using the Qubit High Sensitivity Assay kit (Invitrogen) and DNA integrity was assessed through visualization of bands in a 1.2% agarose gel electrophoresis. Prior to library construction, DNA amount per sample was normalized at 100 ng/μL.
The Genotyping-by-Sequencing (GBS) library was constructed using the standard protocol described by Elshire et al. (2011) and employing theApeK I restriction enzyme. Single-end 100-bp sequencing was conducted on the Illumina HiSeq 2500 platform. Samples were demultiplexed according to their respective barcodes. The generated sequences were filtered to remove low-quality sequences and contaminated reads using SeqyClean 1.9.9 (https://github.com/ibest/seqyclean); only high-quality paired-end sequences (with average PhredScore over 24 and over 65bp) were used for further analysis. The SNPs prospection was performed using the software pipeline PyRAD v1.2 (Eaton, 2014). The data was filtered with PLINK. To conduct population genetic analysis, we identified putative loci under selection using software BayeScan 2.1 (Foll & Gaggiotti 2008) using the default values (Q-value <0,05). The LOSITAN software was used to identify loci under selection (Antao et al. 2008) and removed them from the SNPs dataset. The R package r2vcftools (https://github.com/nspope/r2vcftools) - a wrapper for VCFtools (Danecek et al., 2011)⁠ - was used to perform final quality control on the genotype data. Filtering criteria included biallelic SNPs, linkage disequilibrium (r2 < 0.8, (Xuereb et al., 2018)⁠), Hardy-Weinberg Equilibrium (HWE, p > 0.0001), and loci with less than 20% missing data.