We assessed population genetics of 445 S. scabiei mites from 41 host individuals in four host species, including six mites from two dogs from Bakersfield, 137 mites from seven coyotes (five from Bakersfield, one from Fort Irwin, and one from Ventura), 192 mites from 22 kit foxes from Bakersfield, 23 mites from five Taft kit foxes, and 87 mites from four red foxes from Bakersfield and one from Fresno (Table 1). Although occasional loci of particular mites did not amplify (Additional file 1: Table S1), 60 alleles were detected across the 10 microsatellite loci of all four host species, ranging from two alleles detected at SARM-38 to 11 alleles at SARM-33 (Additional file 2: Table S2). A total of 31 private alleles (i.e. alleles found only in one population and among the broader collective populations of study) were detected, distributed among eight loci. All loci showed LD (P = 0.001) and significant deviations from HWE (P < 0.001).
Kit fox mites
Among the Bakersfield kit fox-derived mites, despite relatively large numbers of mites, there was low overall mean allelic richness (R = 1.29) and all except SARM-33, 40, 41 and 44 were monomorphic (Table 2, Additional file 1: Table S1). There were 15 alleles across the variable microsatellite loci including a private allele at SARM-40 at a frequency of 0.003. Values of Ho (0.082) and He (0.117) were relatively low. SARM-33, 40 and 41 were not in HWE, had possible null alleles at frequencies of 0.135, 0.177 and 0.167, respectively, and SARM-33 and 40 were in LD (P = 0.001). Mites collected from Taft kit foxes were similar to Bakersfield kit fox mites and also had low allelic richness (R = 1.19). There were only two polymorphic loci, SARM-40 and 44 (Table 2) and a total of 12 alleles across all loci. No private alleles were detected and, similar to Bakersfield kit fox mites, Ho (0.076) and He (0.085) were low. All alleles were in HWE and there was no LD.
Red fox mites
Red fox-derived mites from Bakersfield had multiple alleles only at the same three loci as kit fox mites, whereas SARM-36 was also variable among the Fresno red fox mites (Table 2). A single private allele was detected at SARM-41 at a frequency of 0.05 in mites from the Fresno red fox and no private alleles were detected in mites from the Bakersfield red foxes. Mean R was low for both Bakersfield and Fresno (R = 1.28) with 15 and 14 total alleles detected in these mite populations, respectively. Values of He from both Bakersfield and Fresno resembled the kit fox mites, as did Ho from Bakersfield, whereas Ho from Fresno was slightly lower (0.114). The only deviation from HWE was at SARM-33 in mites from Bakersfield (P < 0.05). A null allele was detected at SARM-33 at a frequency of 0.295. SARM-33 and SARM-40 were in LD (P = 0.04). Null alleles were not detected in mites from the Fresno red fox and there was no LD.
Domestic dog mites
Despite obtaining only three mites from each of the two dogs, only two loci (SARM-35 and 37) had fixed alleles with a total of 22 alleles detected (Table 2). Two private alleles were detected at SARM-34. Average allelic richness was 2.08 and dog mites had lower Ho (0.017), but higher He (0.368) than mites collected from foxes (Table 1). Only SARM-35 and SARM-37 were in HWE and null alleles were detected at SARM-33 and 34 at frequencies of 0.382 and 0.403, respectively. LD was detected between SARM-33 and 34 (P = 0.003), 36 (P = 0.02), 38 (P = 0.02), 40 (P = 0.03); SARM-34 and 36 (P = 0.05), 38 (P = 0.04) and 40 (P = 0.04); SARM-36 and 38 (P = 0.03) and 40 (P = 0.03); SARM-38 and 40 (P = 0.02); and SARM-41 and 44 (P = 0.003).
Similar to dogs, the majority of loci in mites from coyotes were variable, with only three fixed loci in mites from Ft. Irwin (SARM-36, 37 and 38) and none from Bakersfield (Table 2). However, all except SARM-34 and 44 were monomorphic in mites from Ventura. Total numbers of detectable alleles were much higher than for other mite populations, with 52 detectable alleles in Bakersfield coyotes. There were private alleles in Bakersfield coyotes at eight loci (SARM-33, 34, 35, 37, 40, 41, 44 and 45), a single private allele in Ft. Irwin at SARM-40 and a single private allele in Ventura at SARM-34. Mean R (2.78) was notably high in these Bakersfield coyote mites whereas the lowest overall Ho (0.023) and He (0.032) occurred in the Ventura coyote mites. There was significant HWE departure for Ventura SARM-44 (P = 0.01) and for all loci in Bakersfield coyote mites (P < 0.05), but none in Ft. Irwin. Null alleles were detected at all loci except SARM-45 at frequencies ranging between 0.121 and 0.364 for Bakersfield coyote mites. Bakersfield coyote mites had LD at SARM-33 and 34 (P = 0.002), 36 (P = 0.02), 37 (P = 0.001), 38 (P = 0.001), 40 (P = 0.001), 41 (P = 0.001), 44 (P = 0.001) and 45 (P = 0.02); SARM-34 and 36 (P = 0.001), 38 (P = 0.001) and 44 (P = 0.002); SARM-35 and 45 (P = 0.001); SARM-36 and 38 (P = 0.001); SARM-37 and 38 (P = 0.001), 40 (P = 0.001) and 44 (P = 0.001); and SARM-40 and 44 (P = 0.001) and 45 (P = 0.01). LD was also detected in Ft. Irwin coyote mites between SARM-36 and 40 (P = 0.03).
Host population differentiation
AMOVA analysis showed significant differentiation among host species-derived mite populations (46.8%, P = 0.001). Pairwise FST values showed the most closely related populations were mites from Bakersfield and Taft kit foxes (FST = 0.038), followed by mites from kit foxes and red foxes from Bakersfield (FST = 0.05, Table 3). The least related mites were from the Taft kit foxes and Ventura coyote (FST = 0. 843). Bakersfield kit fox and Bakersfield coyote mites were also genetically distinct (0.508) as were Bakersfield coyote and dog mites (0.168). These relationships were also clear on the scatterplot of the PCA of mites, on which Bakersfield and Taft mites clustered together, but were distinct from all other mites (Fig. 2).
Bayesian clustering of the total data set revealed maximum log Pr (X|K) for three clusters. The first cluster was composed of the mite samples from dogs, Bakersfield coyotes, Santa Barbara coyotes, Ventura coyotes and Fresno red foxes. The second group included some of the mite samples from Bakersfield red foxes and some of the mite samples from Bakersfield kit foxes. The last cluster included the remainder of the Bakersfield red fox and kit fox mite samples, as well as the Taft kit fox mite samples (Fig. 3). The data show strong separation between the group composed of mites collected from dogs, coyotes and Fresno red foxes and a second group composed of mites collected from kit fox and Bakersfield red fox. Additional structure is apparent within the sample of mites collected from kit fox and Bakersfield red fox.