Window-based population genetic analysis
To understand genetic differentiation in monarchs, we calculated various population genetic statistics using Vcftools v. 0.1.15 (Danecek et al. , 2011) for individual populations and for pairwise population comparisons. To reduce the number of false positives, we only considered SNPs that were covered in all individuals in the population for the population-based statistics and SNPs that were covered in all individuals in both populations in pairwise comparison statistics. Nucleotide diversity (θπ ) and Tajima’s D (Td ) were calculated in windows of 10,000 base pairs (10kb) across the genome using Vcftools v. 0.1.15 (Danecek et al. , 2011). Western monarchs were down-sampled to match the number of eastern samples (8 males and 6 females) to calculate Tajima’s D (Td ) and allele frequencies using Vcftools v. 0.1.15 (Danecek et al. , 2011). We calculated genetic differentiation (FST ) for each site using Vcftools v. 0.1.15 (Daneceket al. , 2011) and averaged across the genome in windows of 10kb. To ensure that our conclusions were not driven by genomic window size, we also calculated genetic differentiation (FST ) for different window sizes (100 bp, 500 bp and 5,000 bp) to verify our findings. Absolute divergence (DXY ) was calculated in windows of 10kb across the genome using the allele frequencies. Windows with less than 10% of total sites covered were filtered out to eliminate extremely high values. Fixed, shared and private polymorphisms were calculated between eastern and western monarchs using the allele frequencies. FST values were Z-transformed (FSTZ = (Window FST / Genome AverageFST ) / Standard deviation of Genome wideFST ) to obtain the relative genetic differentiation in the windows to the genomic mean to identify outlier windows. The top 1% of theFSTZ values were selected as the genetic differentiation outliers and the bottom 1% Tajima’s D values were selected as Tajima’s D outliers.