Obtaining genome-wide SNPs in multiple samples
After trimming adapter sequences with Trimmomatic v0.39 (Bolger et al., 2014), only paired-end data were mapped to a reference genome using NextGenMap v0.5.2 (Sedlazeck et al, 2013), and BAM files were generated. After the BAM files were sorted using SAMTools v1.7 (Li, H. et al., 2009), each individual was genotyped and VCF files were generated by HaplotypeCaller in Picard v2.25.1 (Broad Institute, 2021) and GATK v4.2.0.0 (McKenna et al., 2010). Then, only biallelic SNPs were identified and hard filtered (MQ < 30.00, SOR > 4.000, QD < 2.00, FS > 60.000, MQRankSum < –20.000, ReadPosRankSum < –10.000, and ReadPosRankSum > 10.000) with VariantFiltration in GATK. We also determined the sequence of the mitochondrial DNA from F0 individual f12A using GetOrganelle (Jin et al., 2020) and mapped the sequences of all other individuals for this reference and searched for SNPs in the same way as for the whole genome, except that the depth was set to >50 in the filtering conditions for detecting SNPs in mitochondrial DNA.