Genomic DNA was isolated from young leaves of 5–10 seedlings per accession, using the DNeasy 96 Plant Mini Kit (Qiagen, Germany). Genotyping by sequencing (GBS) was performed following the procedure reported by Elshire et al., (2011). Briefly, DNA was digested with the restriction enzyme ApeK I, barcoded libraries were prepared to track each accession, and the DNA sequence corresponding to the region flanking the ApeK I site was obtained on an Illumina HiSeq 2000 platform by LGC Genomics GmbH (Berlin, Germany).Fastq files were evaluated for sequencing quality. Sequences were mapped to the tomato reference genome version 2.50 together with the same subset of SNPs in the re-sequenced accessions from the 150 tomato genome consortium [doi.org/10.1111/tpj.12616], 8 parents from a tomato MAGIC population [doi.org/10.1186/1471-2164-14-791], and 350 accessions from a third re-sequencing initiative [26] as follows. astQC was used to evaluate the quality of the sequenced reads. High-quality reads were mapped against the S. lycopersicum genome build 2.50 using BWA-MEM . After mapping, in order to avoid possible false positives caused by misalignment, the PHRED quality of three aligned nucleotides from each read end was set to 0).
SNP calling was carried out by freebayes using the following parameters: a minimum mapping quality of 57, 5 best alleles, 20 minimum base quality, 0.05 maximum mismatch read alignment rate, 10 minimum coverage, 2 minimum alternate allele count, and 0.2 minimum alternate fraction. To avoid regions with potential assembly problems in the reference genome, the Heinz 1706 reads used to build the reference genome were mapped against the reference assembly version SL2.50. A 50X mean coverage was obtained. Any region having a coverage higher than 200X was removed from the SNP calling.