Assessing genomic differentiation
We generated maximum-likelihood phylogenies for complete mitochondrial
genomes and autosomal SNPs, with and without the inclusion of an
outgroup, using RAxML (GTRCAT model, 10,000 bootstrap replicates, random
starting seed; Stamatakis 2014). It should be noted that phylogeny
inference using highly variable data (e.g., SNPs) can induce acquisition
bias resulting in longer branch lengths (Leaché et al. 2015).
Phylogenies were visualized in FigTree v1.2.2
(http://tree.bio.ed.ac.uk/software/figtree/).
Principal component analyses (PCA) were run and visualized using the
SNPrelate (Zheng et al. 2012) R v3.3.4 library (R Development
Core Team 2008). Diversity statistics (FIS ,
FST, FS or relatedness2 statistics,
nucleotide diversity (π), and Tajima’s D ) were calculated in
VCFtools (Commands: –het, –weir-fst-pop, –relatedness2,
–window-pi, –TajimaD, respectively), with π and Dcalculated in 100 bp window intervals. F2 statistics were
generated from the compute_moment_stats and
compute_most_additive_trees functions in MixMapper (Lipson et
al. 2013) with 1000 bootstraps and SNP blocks of 100 (1 per 100 bp) in
MatLab 2018 (The MathWorks, Inc., Natick, Massachusetts, USA).