3.1 Genetic diversity of the European mink captive populations
The microsatellite markers analysis demonstrated an overall population allelic richness per locus of 2.69 with an average of 2.49 in the western population, and 2.82 in the eastern population. Heterozygosity values were lower in the western population compared to eastern population (Table 1). Bayesian assignment recovered two genetic clusters within our population, and no admixture pattern were detected. All individuals clustered according to populations, corresponding to the two different breeding facilities. The offspring with parents of each population was assigned to the western population according to clustering (p(Kwestern)=0.975). Multilocus heterozygosity was slightly higher in eastern than western populations (Kruskal-Wallis: χ2=3.4761; p-value=0.0623), but the inbreeding coefficient (F) was not (Kruskal-Wallis: χ2=0.085714; p-value=0.7697). Overall, sex and birth location had no significant effect on neutral markers’ diversity and richness. PERMANOVA on genetic distance based on the microsatellite markers detected no variation according to E-mink population, sex and birth location (Table 2).
Raw MHC amplicon sequencing data consisted of 3,084,478 raw reads with an average length of 230 base pairs for MCH-I and 130 base pairs for MHC-II. After processing, we characterized 13 MHC-I motifs (amino acid sequences) and 6 MHC-II motifs. The average number of motifs per individual was 5.3 and 3.08 (range: 2-9; 2-4) for MHC-I and MHC-II genes respectively, indicating the presence of at least five and two copies for the two regions. For the MHC-I gene, three motifs were strictly present in the eastern E-mink, and one motif in the western population. Comparatively, no motifs were unique to eastern E-mink for MHC-IIex2 gene, and three were strictly found in western E-mink (Figure 1). Spearman correlation tests allowed us to detect haplotype blocks for both genes, mostly attributed to the eastern population (with the motifs Mulu:MHC-I*0003, Mulu:MHC-I*0008, Mulu:MHC-I*0012, Mulu:MHC-I*0013 and Mulu:MHC-I*0015 for MHC-I and Mulu:DRB*90701, EU263553 for MHC-II) and western E-mink (Mulu:MHC-I*0007, Mulu:MHC-I*0009 and Mulu:MHC-I*0011 for MHC-I, KM371114_EU263551, EU263558_LC055119, EU263550_EU263557 and EU263554_EU263552_EU263556 for MHC-II, Figure S2 & Table S2). Most of the variation encountered in both genes was expressed in amino acid residues that influence the binding of CD4 and CD8 glycoproteins involved in antigen presentation for adaptive immunity (Figure S3).
Motif richness and divergence (Faith’s PD) were significantly higher in the western population compared to eastern E-mink for MHC-II gene (Kruskal-Wallis: χ2=13.456, p-value=0.0002; χ2=8.0614, p-value=0.0045; respectively). However, for MHC-I, divergence was higher in eastern E-mink compared to the western population, but not motif richness (Kruskal-Wallis: χ2=5.0097, p-value=0.0252; χ2=1.5456, p-value=0.2138, respectively). No changes in motif richness nor divergence were observed according to sex for the two genes. However, we did observe significant variation in MHC-II richness according to birth location (Kruskal-Wallis: χ2=10.854, p-value=0.0125), and a Dunn test with Benjamini-Hochberg correction only detected higher motif richness for the MHC-II gene in captive-born E-mink in Spain compared to the EEP (Dunn: Z=-2.748, adjusted p-value=0.0358). PERMANOVA detected a significant influence of E-mink sex for MHC-I genetic distance, as well an influence of mink population close to the significance threshold (Figure S5), whereas E-mink population was the only variable influenced MHC-II composition variation (Table 2). Finally, Mantel tests showed a positive correlation between MHC class I and neutral markers distances (Mantel: r=0.2761, p-value=0.001).