For mtDNA data, an AMOVA was performed at both the nucleotide and haplotype level. GenAlEx 6.5 was used to estimate FST based on haplotype frequencies (Griffiths et al. 2011). For the nucleotide level analysis, MODELTEST 2.1.1 (Guindon and Gascuel 2003, Posada 2008, Darriba et al. 2012) identified Tamura and Nei (1993), assuming equal base frequencies

with gamma correction (α = 0.12), as the most appropriate model of DNA evolution given the sequence data. Arlequin 3.5 was used to calculate individual pairwise nucleotide distances Decitabine solubility dmso under this model of sequence evolution. In keeping with the common practice in similar studies of humpback whales (Olavarría et al. 2007, Rosenbaum et al. 2009) we use the notation FST for haplotype frequency differentiation and ΦST for nucleotide differentiation (e.g. Weir and Cockerham 1984, Takahata and Palumbi 1985, Hudson et al. 1992). To evaluate the genetic data without the need to impose a priori population structure, we applied the Bayesian clustering approach implemented in the software STRUCTURE version 2.3.1 (Pritchard et al. 2000) to the microsatellite data set. We also repeated the analysis using the three sampling locations as priors to assess the influence of BGB324 clinical trial geography (LocPrior model; Hubisz et al.

2009). This method attempts to partition samples into K group(s) such that the loci in those groups are in Hardy-Weinberg equilibrium, and linkage equilibrium. An ancestry model of admixture and correlated allele frequencies were assumed among populations with 10,000 burn-in steps and 300,000 Markov Chain Monte Carlo repetitions. Five replicates for each number of populations (K = 1 to 6) were performed to verify that the number of populations identified was

consistent between runs. STRUCTURE output was summarized and evaluated using the software CorrSieve (Campana et al. 2011). Potential differences in female and male dispersal rates between eastern and western Australia were investigated using both genetic markers by calculating pairwise estimates check details of FST among populations for each sex. For comparative purposes, Jost’s DEST was also calculated for microsatellite data. DEST was not calculated for mtDNA data as the method is based on differences in interpopulation gene diversity (Jost 2008), and as such, does not take into account the evolutionary relationships between haplotypes (Meirmans and Hedrick 2011). To investigate genetic structure between the Australian populations and those of the South Pacific (including New Caledonia, Tonga, Cook Islands, French Polynesia, and Colombia we combined our mtDNA data with those presented by Olavarría et al. (2007) and calculated FST and ΦST for pairwise comparisons. The correlation between geographic and genetic distances was analyzed using a Mantel test with statistical testing based on 999 random permutations conducted in GenAlEx 6.5 (Smouse et al. 1986, Smouse and Long 1992).