3.1 Genetic structure
The population structure analysis, PCAngsd and NGSAdmix(Figure 2), based on a set of ~1.7 million genotype
likelihoods, indicated that Harbour porpoise subspecies and populations
clustered together. The PCA including all samples (Figure 2A) show two
major axes of differentiation: subspecies (PC1) and populations of the
North Atlantic subspecies (PC2). The first principal axis, explaining
5.8% of the variance, separates the BLS and one IBE individual from the
rest of the samples, suggesting that this IBE individual could belong to
the Iberian subspecies. The second principal component (2.2% explained
variance) divides Baltic Sea porpoises (BES and PBS) from the Atlantic
porpoises (CA, ICE, BAS, NOS), except for one porpoise bycaught in
Latvian waters that clustered with the Atlantic ones. Nine samples (1
NOS, 4 BES and 4PBS) were located between the Atlantic and Baltic
clusters, but were more closely related to the Baltic one (Figure 2A).
When analyzing the PCA without BLS samples (Figure 2C), PC1 (3%
variance explained) also divides the Baltic from the Atlantic samples,
but in this case only 6 individuals were located between the two groups,
as the other three PBS samples were separated by PC2 (2% variance
explained). The PCA including only porpoises from the Baltic region
(Figure S4A) further suggested the presence of three populations in this
small area. The admixture results were consistent with the population
structure identified in the PCAs. On the dataset including BLS samples
(Figure 2B), K2 separated BLS porpoises from the others, K3 subdivided
the Baltic from the Atlantic and K4 isolated the same three PBS samples
from the remainder of the Baltic. The admixture analysis without BLS
(Figure 2D) and only Baltic region samples (Figure S4B) first divided
Baltic samples from the rest (K2), and then the same three PBS
individuals from the remaining of the Baltic (K3).
Based on the genetic structure results (Figure 2A,B), and published
telemetry and passive acoustic data, we chose K=4 as the highest level
of structure and assigned individuals to any of the four clusters if the
likelihood of membership was ≥ 75% in the admixture analysis. 66 out of
the 72 individuals could be assigned to one of the four clusters,
leaving six individuals that were admixed between Atlantic and BES
cluster. This division resulted in 26 porpoises assigned to the Atlantic
population (red cluster), 6 admixed between Atlantic and BES, 32
assigned to the BES population (dark blue cluster), 3 assigned to the
Proper Baltic Sea population (light blue cluster) and 5 assigned to the
BLS subspecies (black cluster). Both IBE samples clustered with the
Atlantic. For the analysis at the population level (Treemix ,SMC++ and population summary statistics), we excluded the 6
specimens admixed between Atlantic and BES, as well as 11 individuals
found in the PBS region, but assigned to the Atlantic (red) or BES (dark
blue) cluster, such that PBS is represented by the three specimens
forming the PBS-specific light blue cluster in Figures 2B and 2D.