A Suite of Genetic Markers Useful in Assessing Wildcat (Felis silvestris ssp.)— Domestic Cat (Felis silvestris catus) Admixture

Abstract

The wildcat (Felis silvestris ssp.) is a conservation concern largely due to introgressive hybridization with its congener F. s. catus, the common domestic cat. Because of a recent divergence and entirely overlapping ranges, hybridization is common and pervasive between these taxa threatening the genetic integrity of remaining wildcat populations. Identifying pure wildcats for inclusion in conservation programs using current morphological discriminants is difficult because of gross similarity between them and the domestic, critically hampering conservation efforts. Here, we present a vetted panel of microsatellite loci and mitochondrial polymorphisms informative for each of the 5 naturally evolved wildcat subspecies and the derived domestic cat. We also present reference genotypes for each assignment class. Together, these marker sets and corresponding reference genotypes allow for the development of a genetic rational for defining “units of conservation” within a phylogenetically based taxonomy of the entire F. silvestris species complex. We anticipate this marker panel will allow conservators to assess genetic integrity and quantify admixture in managed wildcat populations and to be a starting point for more in-depth analysis of hybridization.

The wildcat (Felis silvestris ssp.) is among the most common of wild felids, yet, like the other 37 extant cat species, it faces serious threats to its long-term survival (Nowell et al. 1996; Driscoll and Nowell 2008). In addition to the more typical anthropogenic threats (habitat loss, hunting), the wildcat is under pressure from its overwhelmingly more common domestic derivative the house cat (F. s. catus), primarily in the form of genetic introgression (Ragni and Randi 1986; French et al. 1988; Randi and Ragni 1991; Stahl and Artois 1991; Hubbard et al. 1992; Nowell et al. 1996; Beaumont et al. 2001; Pierpaoli et al. 2003; Macdonald et al. 2004, 2010; Yamaguchi, Driscoll, et al. 2004; Yamaguchi, Kitchener, et al. 2004; Kitchener et al. 2005; Germain et al. 2008; Oliveira et al. 2008; Kitchener and Rees 2009).

The domestic cat is the world’s most popular pet with an estimated 600 million in household association (Legay 1986) and an additional 600 million living independent of humans. Domestic cat morphology and hunting abilities are little changed from the wildcat ancestor. These native abilities enable domestics to compete effectively in natural environments (Biro et al. 2005). Additionally, due to their more flexible and anthropophilic behavior, they also can derive comparatively greater benefit from the human-dominated environments that now constitute 40% of the Earth’s land surface, giving them a significant overall ecological advantage (Macdonald et al. 2000). Feral domestic cats inhabit most sea islands and every continent except Antarctica.

Because domestic cats are ubiquitous, interbreeding with wildcats is pervasive and indeed has been reported in all areas where the problem has been studied (Smithers 1968; Coleman et al. 1997; Driscoll et al. 2007; Sarmento et al. 2009). This unrelenting introgressive effect threatens the genetic integrity of remaining wildcat populations.

Genetic infiltration of domestic alleles into native gene pools may result in the detrimental breakup of locally adaptive autochonous wildcat gene complexes (Rhymer and Simberloff 1996; Allendorf et al. 2001; McGinnity et al. 2003). Hybrid recombinant genotypes are expected to be less fit than either parental type because they have never been tested by natural selection (Turelli et al. 2001). Wildcat populations thus become less suited to their natural environments, more likely to occupy those habitats favored by domestics, and are so pushed along a slippery slope of greater and greater admixture and compromised genetic integrity (Germain et al. 2008). It is also conceivable that selection may actually favor the incorporation of some domestic genes into a wild population (Anderson et al. 2009). In the case of cats, because “domestic” behavior is characterized by a less fearful attitude toward humans (Driscoll, Macdonald, et al. 2009), domestic cats and hybrids are more likely to access resources entailing from human activities than are wildcats (Germain et al. 2008).

To maintain genetic integrity, it is important for conservation, reintroduction, and captive breeding programs to enroll only purebred wildcats. Notably, legal protection is not extended to wildcat/domestic cat hybrids. Similarly, it is important to vet all founders and contributors to conservation programs so as to preserve diversity and ensure the widest possible genetic base in order to forestall inbreeding depression and maintain evolutionary potential. At present, wildcats are identified primarily based on morphology and pelage (Kitchener and Easterbee 1992; Macdonald et al. 2004; Yamaguchi, Driscoll, et al. 2004; Yamaguchi, Kitchener, et al. 2004; Kitchener et al. 2005). However, surveys found a high error rate (39%) among experts asked to distinguish between wildcats and domestics (Ragni 1993), and due to poor management and a lack of documentation, subspecies admixture cannot be excluded from wildcats held in captivity. As a result of these and other factors, conservation has been problematic.

Here, we outline a 2-step protocol for assigning subspecies affiliation of cats of unknown genetic origin based on short tandem repeat (STR) allele frequency variation and subspecies informative mitochondrial DNA (mtDNA) markers ascertained from a comprehensive sampling of both human-associated and wild-living cats. The complete protocol and Supplementary Tables are deposited at DRYAD. We provide “reference” mitochondrial haplotypes (Supplementary Tables 2 and 3) with associated primers (Supplementary Table 4) and reference STR genotypes with associated primers (Supplementary Tables 5 and 6) derived from a world-wide sampling of wild-living and human-dependant cats whose ancestry has been inferred based on corroborated mtDNA and nuclear STR phylogeny, Bayesian cluster analysis of STR genotypes and sample provenance (Driscoll et al. 2007).

Supplementary Table 1 lists provenance, mtDNA phylogenetic clade, and STR phylogenetic grouping for each individual enrolled in the ascertainment study (Driscoll et al. 2007). Select individuals in Supplementary Table 1 are designated as “references.” These are individuals that 1) fell into the appropriate mtDNA clade, 2) fell into the appropriate STR phylogeny clade, 3) was provenanced to an appropriate location, and 4) consistently had a membership coefficient, or Q value, > 0.80 in the appropriate partition according to a STRUCTURE clustering that incorporated no prior population information (for more details of the analysis procedure, see Supplementary Files in Driscoll et al. (2007)). Specifically, these reference individuals were so designated according to the following criteria:

• European Wildcat: mtDNA Clade I; STR clade A; provenanced to Europe; Q value > 0.80 in the “WC” cluster (K = 2) and Q value > 0.80 in the “EWC” cluster (K = 3–7)

• Southern African Wildcat: mtDNA Clade II; STR clade B; provenanced to southern Africa; Q value > 0.80 in the “WC” cluster (K = 2) and Q value > 0.80 in the “AfWC” cluster (K = 7).

• Asian Wildcat: mtDNA Clade III ONLY; STR clade C; provenanced to central Asia; Q value > 0.80 in the “WC” cluster (K = 2) and Q value > 0.80 in the “AWC” cluster (K = 7).

• Western domestic cats: mtDNA Clade IV; STR clade D; provenanced as a breed cat, UK house cat or French village cat; Q value > 0.80 in the “DC” cluster (K = 2) and Q value > 0.80 in an “EDC” cluster from K = 3 and thereafter.

• Eastern domestic cat: mtDNA Clade IV; STR clade D; provenanced as a breed cat, Japanese village cat, UK housecat, or French village cat (to allow for exotics as pets, e.g., Siamese); Q value > 0.80 in the “DC” cluster (K = 2) and Q value > 0.80 the “ADC” cluster from K = 4 and thereafter.

Mixed domestic cats are those that showed evidence of both eastern and western genetic influence but were conclusively pure domestics.

For greatest efficiency, ascertaining ancestry of unknown individuals should start with mitochondrial assessment followed by STR genotyping. Mitochondria’s indeliquescent phylogenetic signature can potentially detect distant admixture events. Only after matrilineal continuity for the lineage of concern is established need the potentially more costly STR typing be done. See Supplementary Information for amplification conditions.

To begin, an initial assessment by sequencing of informative mtDNA segments (Supplementary Tables 2 and 3) will establish maternal ancestry to 1 of 5 naturally evolved wildcat subspecies lineages. In such an analysis, domestic cats (F. silvestris catus) will share assignment with F. silvestris lybica, the wildcat subspecies of north Africa and the Near East from which domestic cats are uniquely derived (Driscoll et al. 2007). Thus, for example, in analysis of European wildcat populations, mtDNA matrilines that cluster with F. s catus/lybica can be presumed to represent admixture with domestic cats as there is no evidence for the translocation and release to the European wild of wild-type F. s. lybica.

The second step is Bayesian clustering analysis of microsatellite genotypes in comparison with genotypes from well vetted “reference” individuals (reference genotypes are provided in Supplementary Table 4). Supplementary Table 5 provides primer sequences and chromosomal location for each of 36 STR loci used in analysis of the reference individuals. Allele frequency per locus is given for each assignment category (including both STR and mtDNA subspecies assignments). Using the reference individuals listed in Supplementary Table 1 as “learning samples” in STRUCTURE (Pritchard et al. 2000; Pritchard and Wen 2004) analysis provides a baseline comparison for each of the subspecies allowing reliable diagnosis of individuals of unknown ancestry as well as estimates of admixture. In practice, the level of confidence in an individual’s ancestry can be adjusted according to criteria established for each conservation application. Thus, for example, where a “loose” definition of purity is acceptable, a Q value of 0.8 might be used, whereas a more stringent definition of pure might use a value of 0.9 or greater. Membership coefficients of individuals that are highly skewed toward 1 or 2 clusters suggest STRUCTURE is detecting genuine population structure (Rosenberg et al. 2002). Individuals with assignment to 2 or more groups, or those with discordant mtDNA/STR profiles, are considered admixed.

Individuals that have only STR data should be considered as incomplete for admixture analysis, regardless of the calculated Q value. Because of the long history of domestic/wildcat introgression (Bewick 1807; Driscoll, Clutton-Brock, et al. 2009; Driscoll et al. 2007; Suminski 1962), it is possible that the distant descendants of a female domestic/male wildcat breeding will appear “pure” wildcat in analysis of nuclear markers (i.e., STRs) yet will still carry the maternally inherited mitochondria introduced by the initial cross.

Funding

National Institutes of Health.