3.4 Neutral - adaptive population structure and genetic differentiation
Neutral and adaptive SNP datasets revealed a similar population structure (Figure 6, S14). On the dataset without BLS, both PCAs of the neutral (Figure 6A) and adaptive SNPs identified with Pcadapt(Figure 6B) separated the three North Atlantic populations, i.e., Atlantic, BES and PBS. The only difference was that in the neutral set, the three PBS porpoises were separated by PC2 (1.87%) and in the adaptive set by PC1 (15.4%). As to the dataset with BLS, the PCA with the set of neutral SNPs (Figure S14A) separated BLS individuals (PC1) and Baltic from Atlantic porpoises (PC2). The PCA inferred with the set of adaptive SNPs (Figure S14B) was however slightly different: PC1 also separated BLS porpoises from the rest, but PC2 divided the three PBS samples (light blue cluster in Figure 2) from the rest.
Genome-wide pairwise Fst levels on the neutral dataset were moderately low (Figure 6C), ranging from ~ 0 to 0.13.Fst levels on the adaptive dataset were higher, ranging from ~ 0 to 0.37. Among Atlantic sampling locations (CA, ICE, BAS and NOS), the adaptive dataset presented slightly lower values than the neutral dataset. However, among Baltic (BES and PBS) populations and between Baltic and Atlantic locations, Fst was higher in the adaptive dataset than in the neutral dataset. The Mantel tests analyzing associations of pairwise-Fst and distance (Figure S14C) among sampling locations were not significant, neither for the neutral nor for the adaptive dataset, as could be expected from the inferred genetic homogeneity across the large geographic range of the open North Atlantic.