3 | RESULTS
3.1 | Genetic diversity and inbreeding
Fine-scale representations of the distribution of ROHs across the longest scaffold analyzed suggest that such tracts are longer and more abundant in threatened S. catenatus individuals relative to non-threatened S. tergeminus individuals (Figure 1). Further analyses confirmed that S. catenatus shows low genetic diversity and high inbreeding levels with respect to S. tergeminus at the genome-wide level (Figure 2). Specifically, when averaged across all individuals, S. catenatus had a mean ΘW (scaled to 103 hereafter) of 2.71 (SE = 0.01) andF ROH of 0.557 (SE = 0.013), while S. tergeminus had a mean ΘW of 4.44 (SE = 0.02) andF ROH of 0.067 (SE = 0.001) (Figure 2a); differences between species for both metrics were highly significant (P < 0.001). These patterns are consistent with repeated reductions in effective population size and higher levels of inbreeding in S. catenatus compared to S. tergeminus .
A comparison of adjusted F ROH values across studies suggests that S. catenatus is moderately inbred with respect to other threatened and endangered species and subspecies (Table S2 – see also Brüniche-Olsen et al., 2018). For these comparisons we recalculated F ROH using data from our study to be comparable with the different size criteria used in different studies (see Table S2). We also recognize that such comparisons are problematic because of varying criteria used to define ROH tracts (e.g., maximum number of heterozygous genotypes allowed to account for sequencing errors), However, based on these adjusted comparisons, S. catenatus is half as inbred as Apennine brown bears (Ursus arctos marsicanus ) that have existed in isolation for hundreds of years (Bennazo et al., 2017), but shows similar levels of inbreeding to Alpine ibex (Capra ibex ) that have recently undergone serial population bottlenecks (Grossen et al., 2020) (Table S2). Our interpretation thatS. catenatus is moderately inbred is also consistent withF ROH values generated from 100 mammals representing species with varying population sizes (van de Valk et al., 2019). Based on a size criterion of ROH > 0.1 Mb, the estimated F ROH value for S. catenatus(0.552) is in the upper third of values for species and subspecies analyzed. On the other hand, F ROH values forS. tergeminus are always representative of an outbred species that shows little inbreeding (Table S2).
In contrast, to species-level comparisons of ΘW andF ROH across the nine S. catenatuspopulations examined show few significant differences for both metrics (Figures 2a and S1). Population-level values of ΘWranged from 2.58 (SE = 0.05) in ROME to 2.78 (SE = 0.01) in KPWA, with only two pairwise comparisons among populations being significant (P < 0.05; Figures 2a and S1). Likewise, values ofF ROH ranged from 0.513 (SE = 0.045) in PROF to 0.608 (SE = 0.033) in CEBO, with only three pairwise comparisons among populations being significant (P < 0.05; Figures 2a and S1). Overall, these results suggest a remarkable degree of uniformity across S. catenatus populations with respect to their long-term demographic trajectories.
S. catenatus genomes were also composed of a substantially greater number of ROHs (N ROH = 364, SE = 10) relative to S. tergeminus genomes (N ROH = 150, SE = 8) (P < 0.001; Figure 2b). This two-fold increase in N ROH (combined with the eight-fold increase in F ROH as above) is consistent with the effects of recent bottlenecks leading to inbreeding in S. catenatus populations, as it implies that ROH tracts have increased in mean length (Figure 2b; also see Figure 1 in Ceballos et al. [2018]).
We also found substantial variation among individuals for bothF ROH and N ROH metrics. For example, there is a two-fold difference in F ROHand even greater variation in N ROH among individuals from the SPVA population (Figure S2). Some of this variation could be due to differences in the degree of relatedness among individuals within single populations, since, unlike the only S. tergeminus population analyzed, S. catenatus populations presented small but perhaps significant (mean = 0.082, SE = 0.012) proportions of close relatives (r xy> 0.125; Figure 3). These patterns are also consistent with a recent onset of inbreeding in S. catenatus populations.