Genetic structure of neutral microsatellites and the NR1D1 cTNR
We first calculated pairwise estimates of Jost’s D at both theNR1D1 locus and our 14 neutral microsatellite loci with the software package diveRsity v1.9.90 (Keenan, McGinnity, Cross, Crozier, & Prodöhl, 2013) implemented in R (R Core Team, 2016). We assigned individuals to populations based on previously reported STRUCTURE results [Newfoundland (Row et al., 2012), Quebec lynx located south of the St. Lawrence River (Koen, Bowman, & Wilson, 2015), Cape Breton Island (Prentice et al., 2017b), and mainland Canada (the remaining distribution of lynx across continental Canada)]. To estimate whether pairwise genetic differentiation at the NR1D1 locus fell within or outside the expected range of neutrality, we plotted pairwise Jost’s D values estimated for our neutral marker dataset with standard error of the mean intervals, and then overlaid each pairwise estimate of Jost’s D for the NR1D1 locus. We used Jost’s D here as it uses the effective number of alleles (Kimura & Crow, 1964), which is a more appropriate approach for outlier-based assessments compared to FST and its analogues, which are dependent on, and can be biased by, within-population heterozygosity (Hedrick, 1999). Further, Jost’s D has been shown to perform better at estimating the differentiation in allele frequencies among populations, whereas FST is preferential for describing the effect of demography on genetic variation (Meirmans & Hedrick, 2011).