Content uploaded by Oliver Berry
Author content
All content in this area was uploaded by Oliver Berry on Jan 22, 2016
Content may be subject to copyright.
Death by sex in an Australian icon: a continent-wide survey reveals
extensive hybridisation between dingoes and domestic dogs
Danielle Stephens1, Alan N. Wilton2, Peter J.S. Fleming3,4, Oliver Berry1,5
1. School of Animal Biology and Invasive Animals Cooperative Research Centre, M092,
The University of Western Australia, Crawley, Western Australia, 6009, Australia.
2. School of Biotechnology and Biomolecular Sciences and Clive and Vera Ramaciotti
Centre for Gene Function Analysis, University of New South Wales, New South Wales
2052, Australia.
3. Vertebrate Pest Research Unit, Biosecurity NSW, NSW Department of Primary
Industries, Orange Agricultural Institute, New South Wales, 2800, Australia.
4. School of Environmental and Rural Sciences, University of New England, Armidale, New
South Wales, 2351, Australia.
5. Current address: CSIRO Oceans and Atmosphere, PMB 5, Floreat, Western Australia,
Australia, 6014.
Keywords: Dingo, Canis familiaris dingo, Canis familiaris, hybridisation, admixture,
conservation, wild dog
1
2
3
4
Abstract
Hybridisation between domesticated animals and their wild counterparts can disrupt adaptive
gene combinations, reduce genetic diversity, extinguish wild populations, and change
ecosystem function. The dingo is a free-ranging dog that is an iconic apex predator and
distributed throughout most of mainland Australia. Dingoes readily hybridise with domestic
dogs, and in many Australian jurisdictions distinct management strategies are dictated by
hybrid status. Yet, the magnitude and spatial extent of domestic dog-dingo hybridisation is
poorly characterised. To address this, we performed a continent-wide analysis of
hybridisation throughout Australia based on 24 locus microsatellite DNA genotypes from
3,637 free-ranging dogs. Although 46% of all free-ranging dogs were classified as pure
dingoes, all regions exhibited some hybridisation, and the magnitude varied substantially. The
south-east of Australia was highly admixed, with 99% of animals being hybrids or feral
domestic dogs, whereas only 13% of the animals from remote central Australia were hybrids.
Almost all free-ranging dogs had some dingo ancestry, indicating that domestic dogs could
have poor survivorship in non-urban Australian environments. Overall, wild pure dingoes
remain the dominant predator over most of Australia, but the speed and extent to which
hybridisation has occurred in the approximately 220 years since the first introduction of
domestic dogs indicates that the process may soon threaten the persistence of pure dingoes.
2
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Introduction
A common consequence of the expansion of human settlements is the interaction between
domestic and wild animals (Stronen et al. 2012 ). Gene flow from domestic animals to their
wild counterparts, through hybridisation, can threaten the integrity of wild lineages in several
ways, including; disrupting adaptive gene complexes, reducing reproductive opportunities,
and genomic swamping of small wild populations (Rhymer & Simberloff 1996). Globally,
rates of hybridisation are increasing (Allendorf et al. 2001), yet, although threats from
hybridisation are well-recognised, the sometimes morphologically cryptic nature of
hybridisation can make it a difficult process to identify and manage effectively (Allendorf et
al. 2001).
In Australia, there are large and widely distributed populations of free-ranging dogs,
including the dingo (Canis lupus dingo), domestic dogs, and their hybrids. Dingoes are
medium-sized, generalist predators (~15 kg adult weight) that have been, or are currently,
present in most regions of mainland Australia (Fleming et al. 2014 ). Dingoes are an ancient
breed of dog that form a discrete clade within the most basal domestic dog lineage (vonHoldt
et al. 2010), and were likely transported to Australia from East Asia approximately 5,000
years ago by seafarers via the south-east Asian archipelago (Gollan 1984; Savolainen et al.
2004). Dingoes have replaced the extinct Tasmanian tiger (Thylacinus cynocephalus) as
Australia’s mainland apex predator, and are thought to fulfil important ecosystem roles by
moderating densities of native herbivores (Letnic and Crowther 2013; Pople et al. 2000 ) and
possibly introduced mesocarnivores (Johnson et al. 2007; Brook et al. 2012 ; Ritchie et al.
2012). Dingoes are designated ‘native’ animals and subject to conservation efforts in some
jurisdictions (Fleming et al. 2001, 2014).
Domestic dogs of multiple breeds have been present in Australia since the arrival of
European settlers in 1788 (Fleming et al. 2001), and interbreeding between domestic dogs
3
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
and dingoes has been observed both directly and through the identification of hybrid
offspring by skull morphometrics and DNA testing (Elledge et al. 2008; Newsome & Corbett
1982, 1985). These studies in several geographic regions suggest hybridisation between
domestic dogs and dingoes has been rapid and that the regional hybrid proportions
correspond with duration of European habitation. The high number of domestic dogs in
Australia and the regular contact between these and dingoes facilitate the potential swamping
of the dingo’s gene pool so that pure dingoes may soon no longer exist in the wild (Daniels &
Corbett 2003). This could lead to the loss of a relict taxon that has a unique position in
Australian ecosystems and could lead to new ecological roles (Claridge et al. 2014).
In addition to the potential loss of an iconic taxon, hybridisation can lead to the
disruption of adaptive gene combinations and a decrease in fitness (Tufto 2010). The
domestication of dogs has proportionally reduced their brain size and introduced other
differences in skull morphology between dingoes and dogs (Newsome et al. 1980). These
differences, which are associated with hearing and mandibular strength, might reduce
survival of hybrids. Alternatively, given the likely few progenitors of the dingoes extant at the
arrival of Europeans (Savolainen et al. 2004), hybridisation may enhance survival by
increasing genetic diversity (Seehausen 2004). The effect that hybridisation might have on
the behavioural ecology of dingoes and hybrids remains unexplored but is of great interest
because of the dingo’s putative role as a trophic regulator (Claridge & Hunt 2008; Claridge et
al. 2010, Glen 2010). A reliable test for the introgression of domestic dog genes into dingo
populations is required before investigation of these effects.
The dingo is also an excellent case study in the ability to detect complicated
hybridisation patterns in a widespread terrestrial mammal. Although hybridisation has been
reported in other wolf-like canids, such as between wolves (Canis lupus and ssp.) and dogs
(C. familiaris: Randi & Lucchini 2002; Muñoz-Fuentes et al. 2010), and between wolves and
4
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
coyotes (Canis latrans; Koblmüller et al. 2009; Wilson et al. 2009), it has usually been at low
levels or geographically restricted (Stronen et al. 2012 ). The regular contact between dingoes
and dogs, the presence of multiple contact sites and often dense populations of free-roaming
domestic dogs, the difficulty in confidently identifying pure dingoes for a genetic ‘baseline’
(e.g., Newsome et al. 1980), the relatively low level of genetic difference between dingoes
and other dogs (compared to interspecific hybrids), the acknowledged use of dingo genes in
at least one established domestic breed (the Australian Cattle Dog or Blue Heeler; Arnstein et
al. 1964) and the likelihood of multiple generations of backcrossing and hybrid interbreeding
are all challenges to predicting accurately the proportion of dingo ancestry in wild dogs.
The extent of domestic dog-dingo hybridisation is poorly known throughout most of
their range, although it is an important consideration for both dingo conservation and
management of free-ranging dogs for livestock protection (Fleming et al. 2006 ; Fleming et
al. In press ). All Australian States and Territories have legislation and policies for the
management and conservation of dingoes (Fleming et al. 2001), yet implementation is
hampered by lack of knowledge about the distribution and abundance of pure dingoes. In
some regions, the identification of a free-ranging dog as a hybrid, domestic dog or a pure
dingo can mean the difference between mandated destruction and active conservation.
Historically in Australia, dingoes and dingo hybrids were identified through
comparison of the skull morphometrics of remotely sourced dingoes with those of known
domestic dogs and captive bred hybrids (Newsome et al. 1980; Newsome & Corbett 1982,
1985). However, this method cannot be used on either live or juvenile animals, and fails to
accurately identify backcrossed animals (Wilton et al. 1999; Daniels & Corbett 2003; Elledge
et al. 2006). Morphometric measurements are also affected by non-heritable factors, such as
levels of nutrition during growth (Newsome & Corbett 1982). Additionally, both the dingo
and domestic dog samples used in the initial comparisons were assumed not to be hybrids
5
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
because the dingoes were acquired from remote central Australian sources and the domestic
dogs were obtained from an urban dog pound 200 km away (Newsome et al. 1980). Although
it would appear reasonable that these populations were isolated and that pure dingoes were
more likely to be extant in remote areas, there is no surety that either group excluded hybrids.
Genetic analysis offers a more direct means to identify hybridisation. Suitable microsatellite
genetic markers have been developed for the purpose of dingo hybrid testing (Wilton et al.
1999; Wilton 2001), and have been successfully applied on relatively small spatial scales
(Elledge et al. 2008; Robley et al. 2010; Newsome et al. 2013).
This study is the first continent-wide survey to estimate the extent of hybridisation in
an animal, and the first extensive survey of Australian free-ranging dogs to identify pure
dingoes. We focus in particular on the western third of the continent, which had been little
studied compared with south-eastern and central Australia. We expected (after Newsome et
al. 1980 and Corbett 2001) the majority of south-eastern Australia to have high levels of
hybridisation and central Australia to have lower levels, in line with previous regional studies
of skull morphometrics (e.g. Newsome & Corbett 1985; Woodall et al. 1996). We employ
both Bayesian clustering and log-of-odds methods for identification of hybrids. We also
evaluate the performance of these methods against a comprehensive set of simulated hybrid
crosses.
6
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
Methods
Sample collection and laboratory analysis
Tissue samples, mostly ear tips or biopsies, of approximately 5 mm2 were collected between
2007 and 2009 from free-ranging dogs culled for livestock protection by land managers or
trapped for behavioural research (e.g. movement behaviour, Claridge et al. 2009; Newsome
et al. 2013). Samples were stored at room temperature in 3 ml of lysis buffer (Longmire et al.
1997). Map coordinates or descriptions of the collection locality were recorded for each
specimen. Specimens of free-roaming dogs (n= 3,941; Figure 1) were obtained from across
every mainland Australian state, representing all of the bioregions in which they occur (West
2008) and comprising the largest and broadest sampling of dingoes and other free-roaming
dogs to date. In addition, samples were collected from 92 owned domestic dogs of multiple
breeds and their crosses, the majority (n = 86) from Yuendumu (22.253° S, 131.801° E) in
central Australia (see Newsome et al. 2013). The owned dogs were used for creation of a
posteriori reference populations (see details below), but were not included in spatial
interpolation of the extent of hybridisation because they were not part of wild-living dog
populations.
DNA was extracted using a manual glass-fibre method (Ivanova et al. 2006) on 96-well
plates. A 1/3 dilution of extracts was used for amplification to mitigate oversized fluorescence
peaks. Twenty-four microsatellite loci were amplified in 5 multiplex PCR reactions. Primer
combinations are listed in Table 1. PCRs were carried out in 10 μl, consisting of: 5 μl Qiagen
Multiplex PCR solution (Qiagen Inc. Valencia, CA, USA), 1 μl Qiagen Q-Solution, 1 μl
DNA, 0.2 μM of each primer and DNAase/RNAase-free water. PCRs were run with 15
minutes at 95 oC for polymerase activation, followed by 35 cycles of 30 seconds at 94 °C, 90
s at 60 °C and 60 s at 72 °C, with 30 minutes final extension at 60 °C. Fragments were run on
an ABI 3730 capillary sequencer, and analysed using GeneMarker® software (SoftGenetics,
7
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
LLC.). To determine the level of error in genotyping, 48 randomly selected specimens were
tested twice to create consensus genotypes, and the error rate calculated using the software
Microchecker (Van Oosterhout et al. 2004).
‘Average 3Q’ tests of dingo purity
All specimens were analysed for dingo purity with the ‘average 3Q’ method of
Wilton (2001) and Wilton et al. 1999. Categorisation of dingo ancestry relied on a
reference sample set of allele frequencies and private alleles from 90 domestic dogs
acquired from urban populations and held by one of the authors (ANW) and 60 dingoes
collected from central Australia (South Australia and Northern Territory) determined pure
from their pedigrees, which were derived from skull morphometrics and appearance of
their progenitors using the methods of Newsome et al. (1980), and Newsome and Corbett
(1982, 1985) and previously tested pure by Wilton et al. (1999) using 14 markers. New
genotypes were compared to the respective dingo and domestic dog reference allele
frequencies to establish the relative probability that a dog is from pure dingo ancestry and
not that of a ¾ dingo. This comparison provides an ‘average 3Q score’, the log of the
probability ratio (log(Pdingo/P3/4 dingo)) (Elledge et al. 2008; Wilton et al. 1999). The
probability was then averaged across loci for comparison across samples with different
numbers of successfully scored loci. Other hybrid crosses were further inferred from the
reference alleles, e.g., a 50% dingo DNA profile was created by averaging the frequency
from domestic dogs and dingoes for each allele (described below). Individuals were then
assigned to a category most likely to represent their percentage of dingo ancestry.
The presence of alleles diagnostic for domestic dog ancestry (‘doglike’ alleles) was
used in conjunction with the 3Q scores to assign individuals to one of 7 categories (as per
Elledge et al. 2008). The categories used were: (1) Pure dingo (3Q score > 0.1, no doglike
8
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
alleles); (2) probable pure dingo (0.05 < 3Q < 0.1, no doglike alleles); (3) probable hybrid,
>75% dingo (0 < 3Q < 0.05), ≥1 doglike alleles); (4) 65-75% dingo (-0.1 < 3Q < 0) , ≥1
doglike alleles; (5) 50-65% dingo (-0.25 < 3Q < -0.1) , ≥1 doglike alleles; (6) less than 50%
dingo (-0.5 < 3Q < -0.25), , ≥1 doglike alleles, and; (7) domestic dog (3Q < -0.5) , ≥1 doglike
alleles. As the 90 domestic dogs and 60 dingoes used for reference in this method were
required to develop the test, we refer to them as ‘a priori’ reference populations.
The fragment analysis equipment used in this study was different from that used to
analyse the fragment data from the original reference groups. Because different equipment
and choice of fluorophores for each locus can cause size differences in microsatellite data
(Pasqualotto et al. 2007), we adjusted microsatellite sizes to match the reference groups of
Wilton (2001) using a combination of running 20 samples on both fragment analysers to
compare size differences, and matching the pattern of allele frequencies between the data
sets. If a locus could not be matched to the a priori reference group (e.g., inconsistent sizes
when both samples were run on each fragment analyser) it was discarded.
Bayesian clustering tests of dingo purity
Although the reference sample-based method of Wilton (2001) has been criticised because of
uncertainty whether dingo reference samples are genuinely free of domestic dog DNA
(Daniels & Corbett 2003), the level of such impurities is likely very low because large allele
frequency differences exist between the reference groups (Wilton et al. 1999). However, the
amount of domestic dog ancestry in hybrids might be underestimated, consequently
overemphasising the dingo ancestry and biasing towards a more optimistic understanding of
the status of dingo hybridisation. Although unlikely in our case given the history of
acquisition and distribution of animals used to obtain the reference collection, a further
potential problem is that the dingo reference samples might not adequately represent the
9
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
extent of genetic diversity throughout the Australian continent (Elledge et al. 2006; Bray et
al. 2009).
To overcome this potential bias, an alternative method, less reliant on a priori
reference samples, using a model-based clustering algorithm that seeks to identify long-
interbreeding groups of individuals can be used. Such methods are widely used to detect
population structure, but have also been applied to characterise admixture, for example
between domestic cats (Felis catus) and wildcats (F. silvestris) in Europe (Beaumont et al.
2001), and red wolves (Canis rufus) and coyotes (Canis latrans; Bohling & Waits 2011;
Bohling et al. 2013). We used the program Structure (v2.3.3; Pritchard et al. 2000) to
estimate the contribution of dingo and domestic dog ancestry to each individual’s genome.
Structure clusters individuals in order to maximise conformance to Hardy-Weinberg and
linkage equilibria, thereby identifying clusters of interbreeding individuals, and attributes all
or part of each individual’s genome to a pre-defined, but adjustable, number of clusters
(Pritchard et al. 2000). Each analysis performed in this study was run for 200,000 iterations
with 20,000 burn-in runs, which was sufficient for the parameters to reach convergence,
using the admixture and correlated allele frequency models (Falush et al. 2003). In addition,
10 replicates were performed for the selected value of K and q values averaged with the
program CLUMPP (Jakobsson & Rosenberg 2007).
The Bayesian framework underlying Structure permits prior identity information,
such as the sampling location, to assist in modelling. Moderate population structure exists
within free-ranging dingo populations (Fst ~ 0.05; Stephens 2012), but is significantly weaker
than that between dingoes and domestic dogs (Fst ~ 0.35; see results). In order to focus
clustering on subdivision between dogs and dingoes without influence of geographic
subdivision we employed a ‘learning samples’ approach (Pritchard et al. 2010), where we
identified a posteriori domestic dog and dingo reference samples to establish two reference
10
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
clusters, which provide allele frequency estimates to assign the unknown samples to either
cluster. These new reference individuals were sampled for this purpose. First, all specimens
identified by the average 3Q method as category 1 (pure dingo; n = 361) or 7 (domestic dog;
n = 113, including the 86 Yuendumu dogs and 27 domestic dogs caught during the study)
were selected from the rest of the sample, to reduce the test data to a manageable size and
eliminate samples with a low probability of being useful. Principal coordinates analysis
(PCoA) was then used to identify the domestic dogs and dingoes most genetically distinct
from each other and therefore least likely to be admixed. The PCoA was performed on a
genetic distance matrix (described in Smouse & Peakall 1999) between all individuals in
categories 1 and 7, as implemented in Genalex v6.4 (Peakall & Smouse 2006). Although the
a posteriori domestic dog reference sample selected using PCoA is not representative of all
dog breeds, it is likely to represent the types of breeds and crossbreeds that come into contact
with dingoes, which is more suitable for the purpose of this study. The samples were
subsequently screened with Structure for the level of Q-assignment to each of two source
populations (k=2), and only samples with ≥ 95% assignment to the appropriate population
were retained to create two a posteriori reference populations from the most dissimilar
individuals in each group.
The individuals in the a posteriori reference groups were then added to Structure with
the USEPOPINFO flag turned on and updating allele frequencies from the reference
populations. Structure was re-run with all specimens, the USEPOPINFO flag was turned off
for all non-reference specimens, and the cluster number of k = 2 .These settings ensure only
the allelic information from the a posteriori reference groups are used to assign the remainder
of the sample, and does not use data from the non-reference samples to inform the model.
The selected a posteriori reference group for dingoes contained a large proportion of
individuals from the Western third of the Australian continent (Figure 2). To evaluate whether
11
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
population subdivision could influence estimates of admixture we trialled alternative
reference groups for dingoes varying the contribution of individuals from the western region.
Our results indicated that the geographic origins of the dingo reference group had trivial
influence on estimates of admixture (Supplementary Table 1).
Additional analyses
Heterozygosity (Ho and He) statistics and the number of alleles were also calculated from the
a posteriori reference groups to determine any differences in genetic diversity between the
domestic dogs and dingoes with Arlequin v3.5 (Excoffier & Lischer 2010). Related samples
Wilcoxon signed-rank tests were used to determine if the difference between groups was
significant, implemented using PASW statistics 18 for Windows (SPSS Inc.). The level of
genetic subdivision (Fst) between the a posteriori domestic dog and dingo references was
calculated with Genalex, and significance tested with 999 permutations of individuals among
the two groups.
12
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
Spatial patterns of dingo purity
Results of purity tests were geographically mapped using ArcGIS 9.3 (ESRI Inc., Redlands,
CA, USA), including ordinary kriging analysis to interpolate to areas not sampled and to
display multiple specimens that were sampled at the same site. A prediction map was
generated, using the mean purity results when there were multiple points at the same location.
The spherical model in ArcGIS was used with between 2 and 50 neighbours included. Dogs
sampled in captivity at Yuendumu, Northern Territory, were not included in the interpolation.
Simulations of hybrid populations
Because the true ancestry of wild animals is unknown, and therefore the accuracy of the
reference assignments and clusters difficult to evaluate, we simulated hybrid crosses with the
program Hybridlab (Nielsen et al. 2006) to test the performance of both methods in detecting
various hybrid and backcrossed individuals. Fifty randomly selected individuals from each of
the a posteriori reference samples were used as the ‘parent’ populations, and all simulated
hybrid groups also contained 50 individuals. The crosses (detailed in Table 2) were then
analysed using the clustering and average 3Q methods. The settings used in Structure to
analyse the simulated data were the same as described above for the real data, with all the a
posteriori reference samples that were not used in the simulation included as learning
samples. The means and 95% confidence intervals of each cross were calculated, using the
estimate of the proportion of ancestry in the dingo cluster for the clustering method, and the
average 3Q score was used for the average 3Q method. Both were then compared to the
theoretically expected content of dingo ancestry (Table 2). In relation to the a posteriori
analysis, Structure estimates “probability intervals” around each mean q. However, these
estimates are acknowledged as being conservative (e.g. Bohling & Waits 2011, Trigo et al.
2008), whereas simulation and empirical studies demonstrate that mean q values accurately
13
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
estimate admixture proportions for populations as divergent as dogs and dingoes (Vähä &
Primmer 2006). Therefore, we elected to use simulations as a more direct evaluation of
accuracy than considering the probability intervals reported by Structure.
14
282
283
284
285
Results
Distribution of samples
Although sampling was undertaken predominantly in Western Australia, the distribution of
samples (Fig 1) was broadly representative of the density and dispersion of free-ranging
dingoes and other wild dogs across the continent (see Fig 2 in Fleming et al. 2001, Fleming
et al. 2014). After the removal of specimens without spatial coordinates, those with fewer
than 14 loci successfully amplified, and those with evidence of sporadic contamination (more
than 3 alleles at a locus for more than 4 loci), 3,637 wild dog specimens and 88 captive dogs
were available for analysis (Fig. 1). Loci with 3 or more alleles at < 4 loci were treated as
having missing data. Locus FH2175 was removed from the data set because it could not be
reliably scored due to an inconsistent repeat pattern. Locus FH2293 was not used in average
3Q method tests because it could not be consistently matched to the original a priori
reference population alleles. The number of loci used was therefore 22 for the average 3Q
method tests and 23 for the clustering analysis. The error rate determined from the replicated
genotyping was 0.000 (negligible).
Construction of a posteriori reference groups
PCoA revealed that the specimens identified with the average 3Q method as being pure dingo
formed a single, well-defined cluster composed of animals from throughout the sampled
regions of Australia except the far south-east (Fig 2). The horizontal outlier from the dingo
group in Fig. 2 (x = -0.19) is probably separated from the other dingo specimens because
only 15 loci were successfully typed for this individual. Specimens with a coordinate 1 score
of > 0.15 were used as a preliminary group of reference ‘pure’ dingoes. This value captured
the densest cluster of dingoes and provided an adequate sample size but still a conservative
cut-off. Individuals with PCoA scores < -0.5 were used as the a posteriori reference
15
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
population for domestic dogs in construction of simulated hybrid crosses as they were most
dissimilar to pure dingoes. This population consisted of 86 captive animals and 16 dogs
caught in the wild, from Western Australia (10 specimens), Victoria (5) and Queensland (1),
which clustered with the other domestic dogs in the PCoA. The dingo reference group
contained 242 samples from WA, 23 from QLD, 14 from SA, 1 from NSW and 44 from NT
(Fig. 2 inset). After specimens with < 0.95 population assignment in Structure were removed,
the final sample numbers were n = 322 for the dingo reference group (2 from Western
Australia removed), and n = 102 for the dog group.
The expected and observed heterozygosities and the mean number of alleles per locus
from the a posteriori reference groups were all significantly lower in the dingo than the dog
group (Wilcoxon sign rank test, P < 0.001 for all tests). He, Ho and mean number of alleles
for dingoes were 0.47 (± 0.34 S.D.), 0.38 (± 0.28) and 9.23 (± 7.89) respectively, and for
domestic dogs were 0.76 (± 0.19), 0.69 (± 0.17) and 11.78 (± 7.93). Significant genetic
subdivision between these reference groups was recorded (Fst = 0.353, P < 0.001).
Hybridisation levels and spatial distribution of dingoes
Using the average 3Q method, 1,695 (47%) of the wild dogs were ‘pure’ dingoes (category 1
or 2). Analysis using the clustering method showed 1,664 specimens with cluster assignment
≥ 0.90 for dingo ancestry (46% of the sample) (Fig. 3). Assignment of ≥ 0.90 was chosen as
an arbitrary cut-off for the classification of ‘pure’ dingoes using the Bayesian method. The
clustering method showed equal or fewer pure dingoes in all states than the average 3Q
method (Fig 3(b) and (d)), but hybrids from the former results generally were assigned a
higher percentage of dingo ancestry. The highest proportion of pure dingoes was found in the
Northern Territory (87% of individuals by clustering), with intermediate proportions in WA,
SA and QLD, but very few dingoes were detected in NSW and VIC (1% in each of these
16
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
states according to the clustering method). Less than one percent of all free-ranging dog
specimens exhibited no dingo ancestry.
Performance of hybridisation estimation methods using simulated hybrids
Both methods showed strong agreement between the expected purity scores and the average
purity of individuals that was identified from analysis (Fig. 4). The maximum deviation from
the expected mean was 2.5% for the Bayesian clustering method, on cross 6 (mean deviation
over all crosses was 0.95%). All of the expected values were within the 95% CI of observed
results for both methods. Between 2% and 58% of the individuals within each simulated
cross were incorrectly classified using the a priori method (outside of the ranges shown in
Fig. 4a;
x
= 26%, SD=21.35). Using the a posteriori methods 2-26% were outside of the
expected value ± 10% (
x
= 13.67%, SD=9.05). The percentages of misclassifications were
roughly equal above and below the range (6.7% under and 8% over).
17
336
337
338
339
340
341
342
343
344
345
346
347
348
349
Discussion
Distribution of dingoes and hybrids
This study is the first range-wide assessment of hybridisation between dingoes and domestic
dogs, and the largest investigation of hybridisation in any animal. The ubiquity of domestic
dogs in human settlements and extensive livestock husbandry across Australian rangelands
has provided many contact points where hybridisation with dingoes has occurred. Hybrids are
most prevalent in coastal areas (Fig. 3), which have the longest duration of European
settlement and extensive agriculture, and the highest densities of human settlement, and most
opportunities for contact between dogs and dingoes. These more densely populated areas also
have the most intensive and continued culling programs, which could further facilitate
hybridisation if culling encouraged dispersal through social disruption, as suggested by
Wallach et al. (2009). Although dingoes remain in relatively high proportions in the central
and western parts of the Australian continent, no area is free from hybrids. This finding
places the pattern of hybridisation in the second most extreme category under the scheme of
Allendorf et al. (2001) (type 5: widespread introgression), and on course towards the most
extreme category (type 6: anthropogenically mediated complete admixture).
The pattern of hybridisation found in this study accords with previous morphometrics
studies (e.g., Newsome & Corbett 1985; Woodall et al. 1996; Corbett 2001), but provides
much more detail. This is particularly the case in Western Australia, where only one
morphologically-based study has been reported, which found 74% pure dingoes in a region of
the north west (Corbett 2001), and in South Australia, where no studies have been published.
The trend to hybridisation in coastal areas and near human settlement is also shown on the
west coast, but to a lesser extent than found in the southeast of Australia. Although the central
areas in the Northern Territory contained mostly pure dingoes, the specimens were mostly
18
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
sourced near small, remote human settlements, and the purity level may be higher in more
remote areas.
Southeastern Australia exhibited the highest proportion of hybrids, which reflects the
earlier and more intensive European settlement and livestock production (Fleming et al.
2001). A high incidence of hybrids in this area has been reported previously (22/24 free-
ranging dog samples from southeastern New South Wales and northeastern Victoria, which
were analysed using Elledge et al. (2008) classifications; Claridge et al. (2009): and ranging
from 23% pure dingoes on the south coast of New South Wales to 65% dingoes captured in
the Victorian highlands during 1970-79 using the morphometric method; Newsome &
Corbett 1985). The discrepancy between the DNA and morphology-based assessments likely
reflects inaccuracy of morphological assessments as well as an increased proportion of
hybrids since skulls were collected from this area in the 1960s and 1970s. The question of
dingo purity is particularly relevant to the state of Victoria, as the dingo has recently been
declared a threatened species, with requirements for its protection (Anonymous 2007; Glen
2010), whilst landholders are responsible for controlling wild dogs for livestock protection.
Efforts for dingo conservation in Victoria clearly require a comprehensive program to address
the process of hybridisation. Because of the ubiquity of hybrids and difficulty in
distinguishing them on appearance from dingoes (Elledge et al. 2008), such a program would
likely require physical isolation of extant dingo populations.
The higher prevalence of introgression between dingoes and domestic dogs than
between wolves and domestic dogs may be due to the dingo’s history of domestication
followed by feralisation (Fleming et al. 2012), resulting in decreased wariness of human
settlements and food and water resources (Newsome et al. 2013 ), their concentration around
livestock or artificial watering points on grazing land containing working dogs, or due to
higher population densities of free-ranging dogs, including dingoes, than occurs in wolves
19
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
(Koblmüller et al. 2009; Muñoz-Fuentes et al. 2010). Culling of wild dogs is also widespread
in Australia and Wallach et al. (2009), proposed that subsequent reduced population densities
disrupt social groups and might increase the encounter rate between domestic dogs and
dingoes (Corbett 2001; also discussed with respect to wolves in Vilà et al. 2003).
Alternatively, targeted culling of free-ranging dogs closer to human settlements and livestock
production areas could provide a separation between domestic dogs and pure dingoes
inhabiting more remote regions, which could slow introgression.
Distribution of feral domestic dogs
The low incidence of purely domestic dogs in wild populations (<1% of all specimens)
suggests that they have poor survivorship in the wild. The most likely mating between
dingoes and dogs is generally considered to be between a female dingo and a male domestic
dog, due to the difficulty of a domestic female raising a litter outside of captivity (Corbett
2001). The hybrid offspring of owned domestic dogs and dingoes raised in the wild may be
better able to survive and integrate into dingo societies than domestic dogs raised in captivity,
which would have little experience hunting or socialising with dingoes (Vilà & Wayne 1999;
Daniels & Corbett 2003).
As few populations of feral domestic dogs have been reported in Australia (Newsome
and Corbett 1985, Jones 2009), the extent of hybridisation is likely caused either by historical
introgression and subsequent interbreeding of the hybrids, or an ongoing influx of genes from
roaming domestic pets or wild females visiting domestic dogs and subsequent interbreeding
of hybrids. The presence of mostly hybrids with few representatives from either of the parent
populations suggests that free-ranging dogs in some areas may be approaching the status of a
stable hybrid swarm, where the majority of individuals are the product of multiple
20
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
generations of hybrid crosses and backcrosses (Rhymer & Simberloff 1996; Allendorf et al.
2001).
Performance of the methods against simulated data
Both of the purity testing methods identified all simulated hybrid crosses with high accuracy
(Fig. 4). The minimal bias exhibited by the clustering method indicates that it can measure
dingo purity within a few per cent of the average expected value (not accounting for random
inheritance past the F1 generation). For both methods the simulation provide a conservative
measure of accuracy, as in the clustering method 100 of the specimens were removed from
the a posteriori reference groups to create the simulated hybrids, and in the average 3Q
method only the 3Q score was considered, not the inclusion of doglike alleles. This also
affects the calculation of misclassification rates, which are an underestimate of the true
likelihood of misclassification, but provide a useful comparison between the methods.
Although the clustering method still relies on reference samples, as criticised by Daniels and
Corbett (2003), the samples used were selected due to their distinctness by 3 sequentially
applied tests, rather than prior assumptions of purity.
21
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
The a posteriori dingo reference group used in this study is the most geographically
diverse for testing purity to date (Fig 2); whereas the references used to develop canonical
skull scores were selected from central Australia (Newsome et al.1980; Newsome & Corbett
1982), as were the a priori dingo reference group. It should be noted, however, that despite a
western continent bias in the origins of the a posteriori reference groups, and a central
continent bias in the origins of the a priori groups, the patterns of hybridisation and total
number of pure dingoes found are similar, possibly due to the low genetic diversity and
genetic homogeneity within Australian dingoes (Savolainen et al. 2004). Despite this, the
impact of founder effects and genetic drift on isolated areas such as Fraser Island, where
entry of domestic dogs is prohibited to maintain the purity of its dingoes, should be
considered when using any reference method, as the effects of regional allelic variation and
hybridisation may otherwise be confounded.
The DNA-based methods presented here provide a considerable advance in precision
and versatility on morphometric methods, due to their ability to test juveniles and live
animals, and to detect backcrossed hybrids, and for the clustering method that gives a precise
percentage estimate of dingo ancestry. This study also demonstrates the ability to apply the
DNA testing of hybrid individuals on a large scale; it would have been unfeasible to measure
the number of skull samples equal to the number of tissue specimens used in this study.
Implications for the management of wild dogs
Like wolves in Europe and North America, the dingo retains a unique status in Australia as
both an icon of the natural world and an agricultural pest. There are cultural considerations
surrounding the loss of pure dingoes, including their value to indigenous Australian cultures,
tourism, and the perception by many Australians that they are part of the native fauna
22
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
(Fleming et al. 2001). Conflict between cultural and community attitudes, and management
targets (e.g., culling for livestock protection), can be more productively addressed when the
true status of free-ranging dog populations is known.
The impact of dingo hybridisation on the dynamics of ecosystems could be informed
by this study and the application of purity testing. An important aspect to predicting the
success of dingo reintroductions for mesopredator suppression is the effect that hybridisation
has on the ecological roles of free-ranging dogs (Claridge and Hunt 2008).
A decision on whether some degree of mixing between dogs and dingoes will affect
conservation goals, and whether it is possible or necessary to attempt to retain historical
patterns of genetic diversity, should be made prior to the development of management
strategies (Daniels & Corbett 2003; Claridge & Hunt 2008; Rutledge et al. 2010). This
question has also been raised for the preservation of wolves in the Great Lakes Region of
North America, as pre-European admixture between grey wolves and Great Lakes wolves has
been found (Schwartz & Vucetich 2009; Wheeldon & White 2009). The answer to this
question for dingoes is somewhat dependent on whether admixture has affected their fitness
or adaptive potential. Savolainen et al. (2004) detected evidence that a severe bottleneck
occurred in dingoes upon their arrival in Australia, to the extent that perhaps only one family
had arrived. Significantly lower genetic diversity in dingoes than dogs was also found in this
study, on all measures tested. Breeding with dogs, and the associated increase in genetic
diversity, may be beneficial to dingoes' ability to adapt, but whether this is the case requires
careful testing (Barton 2001; Yuri et al. 2009). However, the high number of hybrids,
including backcrosses, found in this study supports the capacity of hybrids to survive and
breed successfully in the wild. Maladaptation or wasted reproductive opportunities therefore
do not appear to be major issues for dingo-dog hybridisation, but whether traits beneficial to
23
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
free-ranging dogs, but undesirable for humans, livestock production and conservation values,
will emerge remains unknown.
The rapid hybridisation between dingoes and domestic dogs at a continental scale
illustrates the importance of establishing baseline genetic data for taxa in the early stages of
admixture, if genetic swamping is likely to become a problem in the future. Once
hybridisation has become widespread it becomes more difficult to reconstruct what
constitutes a 'pure' wild animal, although this study shows that methods can be developed if a
sufficient sample size is available. Additionally, some genetic diversity within the 'pure'
population may be lost by the almost complete admixture in certain regions of the animal's
range, such as appears to be the case in south-eastern Australia. Because the extent of
hybridisation in dingoes is far greater than that of wolves and dogs in Europe (Randi &
Lucchini 2002) or North America (Muñoz-Fuentes et al. 2010) the study of dingoes provides
a unique opportunity to test the impact of large-scale hybridisation on an ecologically
significant top-order predator.
Conclusions and applications
Both of our purity testing methods revealed very similar patterns (Fig. 3), showing
hybridisation mainly concentrated around the more densely inhabited coastal areas and
settlements. This supports previous assertions that duration and the density of settlement
since 1788 are strong predictors of dingo hybridisation levels (Newsome & Corbett 1985;
Woodall et al. 1996). The low genetic diversity found in dingoes may indicate that
conservation efforts would be better targeted at the more remote regions or islands where
successful isolation of dingoes and dogs is more probable, as minimal genetic diversity would
be lost overall. A study of whether dingoes exhibit population structure between regions
would assist in establishing if this is an appropriate strategy.
24
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
The extent and pace of hybridisation found in this study means that preventing gene
flow between dogs and dingoes requires concerted effort (Fleming et al. 2001; Allendorf et
al. 2001). There are three main areas where a better understanding of the pattern of
hybridisation has advantages in wildlife management: establishing the roles of dingoes and
hybrids within ecosystems; understanding the effect that hybridisation may have on livestock
predation; and providing evidence for the formulation of dingo conservation policy
encompassing social attitudes and values. Changes in body size, prey preference and pack
structure through the hybridisation process could all have great ramifications for the status
quo of ecosystems across the Australian continent and require investigation before the dingo
is lost or changed irreversibly.
25
513
514
515
516
517
518
519
520
521
522
523
Acknowledgements
Over 280 people contributed tissue samples for this study. A. Woolnough, B. Davies, T.
Thompson and Y. Hitchen coordinated and encouraged sampling in Western Australia, as well
as providing valuable support and feedback throughout the project. Y. Hitchen also provided
valuable laboratory assistance. P. Thomson and K. Rose provided feedback on project design
and shared their extensive knowledge of dingo ecology. T. Newsome provided the Yuendumu
dog samples used in the dog reference population and G. Ballard provided samples and
assisted with discussions on project design. F. Allendorf, B. Sacks and A. Glen provided
helpful comments on a draft manuscript, as did Lisette Waits and three anonymous reviewers.
Phil Goulding and Mike Nunweek provided advice on GIS analysis. M.S. Johnson provided
invaluable guidance thoughout this project. This study was undertaken while DS was
supported by an Invasive Animals Cooperative Research Centre Balanced Scientist
Scholarship. This manuscript is dedicated to the memory of Alan Wilton, a gentleman and
pioneer of dingo genetics research.
26
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
References
Adams JR, Lucash C, Schutte L, Waits LP (2007) Locating hybrid individuals in the red wolf
(Canis rufus) experimental population area using a spatially targeted sampling
strategy and faecal DNA genotyping. Molecular Ecology 16, 1823-1834.
Allen B, Allen L, Engeman R, Leung LK-P (2013a) Intraguild relationships between
sympatric predators exposed to lethal control: predator manipulation experiments.
Frontiers in Zoology 10, 39.
Allen BL, Fleming PJS, Allen LR, Engeman RM, Ballard G, Leung LK-P (2013b) As clear as
mud: A critical review of evidence for the ecological roles of Australian dingoes.
Biological Conservation 159, 158-174.
Allendorf FW, Leary RF, Spruell P, Wenburg JK (2001) The problems with hybrids: setting
conservation guidelines. Trends in Ecology & Evolution 16, 613-622.
Andersone Z, Lucchini V, Ozoli J (2002) Hybridisation between wolves and dogs in Latvia as
documented using mitochondrial and microsatellite DNA markers. Mammalian
Biology 67, 79-90.
Anonymous (2007) Final recommendation on nomination 789, Canis lupus dingo. Scientific
Advisory Committee, Department of Sustainability and Environment, Melbourne.
Arnstein P, Cohen DH, Meyer KF (1964) Dingo blood improves famous cattle dog. Journal
of the American Veterinary Medical Association 145, 933-936.
Barton N (2001) The role of hybridization in evolution. Molecular Ecology 10, 551-568.
Beaumont M, Barratt EM, Gottelli D, et al. (2001) Genetic diversity and introgression in the
Scottish wildcat. Molecular Ecology 10, 319-336.
Bohling JH, Waits LP (2011) Assessing the prevalence of hybridization between sympatric
Canis species surrounding the red wolf (Canis rufus) recovery area in North Carolina.
Molecular Ecology 20, 2142-2156.
27
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
Bohling JH, Adams JR, Waits LP (2013) Evaluating the ability of Bayesian clustering
methods to detect hybridization and introgression using an empirical red wolf data set.
Molecular Ecology 22, 74-86.
Bray T, Chikhi L, Sheppy A, Bruford M (2009) The population genetic effects of ancestry
and admixture in a subdivided cattle breed. Animal Genetics 40, 393-400.
Brook LA, Johnson CN, Ritchie EG (2012) Effects of predator control on behaviour of an
apex predator and indirect consequences for mesopredator suppression. Journal of
Applied Ecology 49, 1278-1286.
Claridge, Andrew W. (2013) Examining interactions between dingoes (wild dogs) and
mesopredators: the need for caution when interpreting summary data from previously
published work. Australian Mammalogy, -.
Claridge AW, Hunt R (2008) Evaluating the role of the dingo as a trophic regulator:
additional practical suggestions. Ecological Management & Restoration 9, 116-119.
Claridge AW, Mills DJ, Barry SC (2010) Prevalence of threatened native species in canid
scats from coastal and near-coastal landscapes in south-eastern Australia. Australian
Mammalogy 32, 117-126.
Claridge AW, Mills DJ, Hunt R, Jenkins DJ, Bean J (2009) Satellite tracking of wild dogs in
south-eastern mainland Australian forests: implications for management of a
problematic top-order carnivore. Forest Ecology and Management 258, 814-822.
Claridge, A.W., Spencer, R.J., Wilton, A.N., Jenkins, D.J., Dall, D. J. and Lapidge, S.J. (2014)
When is a dingo not a dingo? Hybridisation with domestic dogs. Pp. 151-172 In. (Eds
A.S. Glen and C.R. Dickman) Carnivores of Australia: past, present and future.
(CSIRO Publishing: Collingwood)
Corbett LK (2001) The Dingo in Australia and Asia, Second edn. JB Books, Marleston,
Australia
28
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
Daniels MJ, Corbett L (2003) Redefining introgressed protected mammals: when is a wildcat
a wild cat and a dingo a wild dog? Wildlife Research 30, 213-218.
Dickman CR, Glen AS, Letnic M (2009) Reintroducing the dingo: can Australia's
conservation wastelands be restored? In: Reintroduction of top-order predators (ed.
Hayward MW, Somers M), pp. 238-269. Wiley-Blackwell.
Einum S, Fleming IA (1997) Genetic divergence and interactions in the wild among native,
farmed and hybrid Atlantic salmon. Journal of Fish Biology 50, 634-651.
Elledge AE, Allen LR, Carlsson B, Wilton AN, Leung LK (2008) An evaluation of genetic
analyses, skull morphology and visual appearance for assessing dingo purity:
implications for dingo conservation. Wildlife Research 35, 812-820.
Elledge AE, Leung LKP, Allen LR, Firestone K, Wilton AN (2006) Assessing the taxonomic
status of dingoes Canis familiaris dingo for conservation. Mammal Review 36, 142-
156.
Excoffier L, Lischer HEL (2010) Arlequin suite ver 3.5: a new series of programs to perform
population genetics analyses under Linux and Windows. Molecular Ecology
Resources 10, 564-567.
Falush D, Stephens M, Pritchard JK (2003) Inference of population structure using multilocus
genotype data: linked loci and correlated allele frequencies. Genetics 164, 1567-1587.
Fleming PJS, Corbett LK, Harden RH, Thomson PC (2001) Managing the impacts of dingoes
and other wild dogs. Bureau of Rural Sciences, Canberra.
Fleming PJS, Allen LR, Lapidge SJ, Robley A, Saunders GR, Thomson PC (2006) Strategic
approach to mitigating the impacts of wild canids: proposed activities of the Invasive
Animals Cooperative Research Centre. Australian Journal of Experimental
Agriculture 46, 753-762.
29
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
Fleming, PJS, Allen, BL, Ballard, G-A (2012) Seven considerations about dingoes as
biodiversity engineers: the socioecological niches of dogs in Australia. Australian
Mammalogy 34, 119-131.
Fleming PJS, Allen BL, Allen LR, Ballard G, Bengsen AJ, Gentle MN, McLeod LJ, Meek
PD, Saunders GR (2014) Management of wild canids in Australia: free-ranging dogs
and red foxes. Pp 105-139 In Carnivores of Australia: Past, Present and Future. (Eds
AS Glen, CR Dickman). CSIRO Publishing, Collingwood
Francisco LV, Langston AA, Mellersh CS, Neal CL, Ostrander EA (1996) A class of highly
polymorphic tetranucleotide repeats for canine genetic mapping. Mammalian
Genome: Official Journal of the International Mammalian Genome Society 7, 359-
362.
Fredholm M, Winterø AK (1995) Variation of short tandem repeats within and between
species belonging to the Canidae family. Mammalian Genome 6, 11-18.
Glen AS (2010) Hybridisation between dingoes and domestic dogs: a comment on Jones
(2009). Australian Mammalogy 32, 76-77.
Glen AS, Dickman CR, Soulé ME, Mackey BG (2007) Evaluating the role of the dingo as a
trophic regulator in Australian ecosystems. Austral Ecology 32, 492-501.
Gollan K (1984) The Australian dingo: in the shadow of man. In: Vertebrate zoogeography
and evolution in Australasia (eds. Archer M, Clayton G), pp. 921-927. Hesperian
Press, Australia.
Holmes NG, Dickens HF, Parker HL, et al. (1995) Eighteen canine microsatellites. Animal
Genetics 26, 132a-133.
Holmes NG, Humphreys SJ, Binns MM, et al. (1993) Isolation and characterization of
microsatellites from the canine genome. Animal Genetics 24, 289-292.
30
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
Hutchings JA (1991) The threat of extinction to native populations experiencing spawning
intrusions by cultured Atlantic salmon. Aquaculture 98, 119-132.
Ivanova NV, Dewaard JR, Hebert PDN (2006) An inexpensive, automation-friendly protocol
for recovering high-quality DNA. Molecular Ecology Notes 6, 998-1002.
Jakobsson M, Rosenberg NA (2007) CLUMPP: a cluster matching and permutation program
for dealing with label switching and multimodality in analysis of population structure.
Bioinformatics 23, 1801-1806.
Johnson CN, Isaac JL, Fisher DO (2007) Rarity of a top predator triggers continent-wide
collapse of mammal prey: dingoes and marsupials in Australia. Proceedings of the
Royal Society B: Biological Sciences 274, 341-346.
Johnson CN, VanDerWal J (2009) Evidence that dingoes limit abundance of a mesopredator
in eastern Australian forests. Journal of Applied Ecology 46, 641-646.
Jones E (2009) Hybridisation between the dingo, Canis lupus dingo, and the domestic dog,
Canis lupus familiaris in Victoria: a critical review. Australian Mammalogy 31, 1-7.
Kidd AG, Bowman J, Lesbarrères D, Schulte-Hostedde AI (2009) Hybridization between
escaped domestic and wild American mink (Neovison vison). Molecular Ecology 18,
1175-1186.
Koblmüller S, Nord M, Wayne RK, Leonard JA (2009) Origin and status of the Great Lakes
wolf. Molecular Ecology 18, 2313-2326.
Letnic M, Koch F (2010) Are dingoes a trophic regulator in arid Australia? A comparison of
mammal communities on either side of the dingo fence. Austral Ecology 35, 167-175.
Letnic M, Crowther MS (2013) Patterns in the abundance of kangaroo populations in arid
Australia are consistent with the exploitation ecosystems hypothesis. Oikos 122, 761-
769.
31
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
Longmire JL, Maltbie M, Baker RJ (1997) Use of “lysis buffer” in DNA isolation and its
implications for museum collections. Occasional Papers, The Museum, Texas Tech
University 163, 1–3.
Mellersh CS, Holmes N, Binns M, Sampson J (1994) Dinucleotide repeat polymorphisms at
four canine loci (LEI 003, LEI 007, LEI 008 and LEI 015). Animal Genetics 25, 125-
126.
Mellersh CS, Langston AA, Acland GM, et al. (1997) A linkage map of the canine genome.
Genomics 46, 326-336.
Mooney HA, Cleland EE (2001) The evolutionary impact of invasive species. Proceedings of
the National Academy of Sciences of the United States of America 98, 5446-5451.
Muñoz-Fuentes V, Darimont C, Paquet P, Leonard J (2010) The genetic legacy of extirpation
and re-colonization in Vancouver Island wolves. Conservation Genetics 11, 547-556.
Newsome AE, Corbett LK (1982) The identity of the dingo II.* Hybridization with domestic
dogs in captivity and in the wild. Australian Journal of Zoology 30, 365-374.
Newsome AE, Corbett LK (1985) The identity of the dingo III.* The incidence of dingoes,
dogs and hybrids and their coat colours in remote and settled regions of Australia.
Australian Journal of Zoology 33, 363-375.
Newsome AE, Corbett L, Carpenter S (1980) The identity of the dingo I. Morphological
discriminants of dingo and dog skulls. Australian Journal of Zoology 28, 615-625.
Newsome TM, Stephens D, Ballard GA, Dickman CR and Fleming PJS (2013) Genetic
profile of dingoes (Canis lupus dingo) and free-roaming domestic dogs (C. l.
familaris) in the Tanami Desert. Wildlife Research 40, 196-206
Newsome TM, Ballard G-A, Dickman, CR, Fleming, PJS, Howden, C (2013) Anthropogenic
resource subsidies determine space use by Australian arid zone dingoes: an improved
resource selection modelling approach. PLoS ONE 8, e63931.
32
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
Nielsen EE, Bach LA, Kotlicki P (2006) Hybridlab (version 1.0): a program for generating
simulated hybrids from population samples. Molecular Ecology Notes 6, 971-973.
Oliveira R, Godinho R, Randi E, Alves PC (2008) Hybridization versus conservation: are
domestic cats threatening the genetic integrity of wildcats (Felis silvestris silvestris)
in Iberian Peninsula? Philosophical Transactions of the Royal Society B: Biological
Sciences 363, 2953-2961.
Ostrander EA, Mapa FA, Yee M, Rine J (1995) One hundred and one new simple sequence
repeat-based markers for the canine genome. Mammalian Genome: Official Journal
of the International Mammalian Genome Society 6, 192-195.
Ostrander EA, Sprague GF, Rine J (1993) Identification and characterization of dinucleotide
repeat (CA)n markers for genetic mapping in dog. Genomics 16, 207-213.
Pasqualotto AC, Denning DW, Anderson MJ (2007) A cautionary tale: lack of consistency in
allele sizes between two laboratories for a published multilocus microsatellite typing
system. Journal of Clinical Microbiology 45, 522-528.
Peakall R, Smouse PE (2006) GENALEX 6: genetic analysis in Excel. Population genetic
software for teaching and research. Molecular Ecology Notes 6, 288-295.
Pople, A. R., Grigg, G.C., Cairns, S.C., Beard, L. A. and Alexander, P. (2000) Trends in the
numbers of red kangaroos and emus on either side of the South Australian dingo
fence: evidence for predator regulation? Wildlife Research 27, 269-276.
Primmer CR, Matthews ME (1993) Canine tetranucleotide repeat polymorphism at the VIAS-
D10 locus. Animal Genetics 24, 332-332.
Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using
multilocus genotype data. Genetics 155, 945-959.
Pritchard JK, Wen X, Falush D (2010) Documentation for Structure software: Version 2.3.
Accessed from http://pritch.bsd.uchicago.edu/structure_software/
33
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
Randi E, Lucchini V (2002) Detecting rare introgression of domestic dog genes into wild
wolf (Canis lupus) populations by Bayesian admixture analyses of microsatellite
variation. Conservation Genetics 3, 29-43.
Randi E, Pierpaoli M, Beaumont M, Ragni B, Sforzi A (2001) Genetic identification of wild
and domestic cats (Felis silvestris) and their hybrids using Bayesian clustering
methods. Molecular Biology and Evolution 18, 1679-1693.
Rhymer JM, Simberloff D (1996) Extinction by hybridization and introgression. Annual
Review of Ecology and Systematics 27, 83-109.
Ritchie EG, Elmhagen B, Glen AS, Letnic M, Ludwig G, McDonald RA (2012) Ecosystem
restoration with teeth: what role for predators? Trends in Ecology and Evolution 27,
265-271.
Robley A, Gormley A, Forsyth DM, Wilton AN, Stephens D (2010) Movements and habitat
selection by wild dogs in eastern Victoria. Australian Mammalagy 32, 23-32.
Roy MS, Geffen E, Smith D, Ostrander EA, Wayne RK (1994) Patterns of differentiation and
hybridization in North American wolflike canids, revealed by analysis of
microsatellite loci. Molecular Biololgy and Evolution 11, 553-570.
Rutledge L, Bos K, Pearce R, White B (2010) Genetic and morphometric analysis of
sixteenth century Canis skull fragments: implications for historic eastern and gray
wolf distribution in North America. Conservation Genetics 11, 1273-1281.
Savolainen P, Leitner T, Wilton AN, Matisoo-Smith E, Lundeberg J (2004) A detailed picture
of the origin of the Australian dingo, obtained from the study of mitochondrial DNA.
Proceedings of the National Academy of Sciences of the United States of America 101,
12387-12390.
Schwartz MK, Vucetich JA (2009) Molecules and beyond: assessing the distinctness of the
Great Lakes wolf. Molecular Ecology 18, 2307-2309.
34
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
Seehausen O (2004) Hybridization and adaptive radiation. Trends in Ecology & Evolution 19,
198-207.
Smouse PE, Peakall R (1999) Spatial autocorrelation analysis of individual multiallele and
multilocus genetic structure. Heredity 82, 561-573.
Stephens D (2012) The Molecular Ecology Of Australian Wild Dogs: Hybridisation, Gene
Flow And Genetic Structure At Multiple Geographic Scales. PhD Thesis. School of
Animal Biology, The University of Western Australia. 122pp.
Stronen AV, Tessier N, Jolicoeur H, Paquet PC., Hénault M, Villemure M, Patterson BR,
Sallows T, Goulet G, Lapointe F-J (2012) Canid hybridization: contemporary
evolution in human-modified landscapes. Ecology and Evolution 2, 2128-2140.
Tufto J (2010) Gene flow from domesticated species to wild relatives: migration load in a
model of multivariate selection. Evolution 64, 180-192.
Trigo T, Freitas T, Kunzler G, et al. (2008) Inter species hybridization among Neotropical‐
cats of the genus Leopardus, and evidence for an introgressive hybrid zone between
L. geoffroyi and L. tigrinus in southern Brazil. Molecular Ecology 17, 4317-4333.
Vähä JP, Primmer CR (2006) Efficiency of model based Bayesian methods for detecting‐
hybrid individuals under different hybridization scenarios and with different numbers
of loci. Molecular Ecology 15, 63-72.
Van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P (2004) MICRO-CHECKER:
software for identifying and correcting genotyping errors in microsatellite data.
Molecular Ecology Notes 4, 535-538.
Vilà C, Walker C, Sundqvist AK, et al. (2003) Combined use of maternal, paternal and bi-
parental genetic markers for the identification of wolf-dog hybrids. Heredity 90, 17-
24.
35
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
Vilà C, Wayne RK (1999) Hybridization between wolves and dogs. Conservation Biology 13,
195-198.
vonHoldt BM, Pollinger JP, Lohmueller KE, et al. (2010) Genome-wide SNP and haplotype
analyses reveal a rich history underlying dog domestication. Nature 464, 898-902.
Wallach AD, Ritchie EG, Read J, O'Neill AJ (2009) More than mere numbers: the impact of
lethal control on the social stability of a top-order predator. PLoS One 4, e6861.
West, P. (2008) Assessing invasive animals in Australia 2008. (National Land & Water
Resources Audit and Invasive Animals Cooperative Research Centre: Canberra)
Wheeldon T, White BN (2009) Genetic analysis of historic western Great Lakes region wolf
samples reveals early Canis lupus/lycaon hybridization. Biology Letters 5, 101-104.
Wilson PJ, Grewal SK, Mallory FF, White BN (2009) Genetic characterization of hybrid
wolves across Ontario. Journal of Heredity 100, S80-S89.
Wilton AN (2001) DNA methods of assessing dingo purity. In: A symposium on the dingo
(eds. Dickman C, Lunney D), pp. 49-56. Royal Zoological Society of New South
Wales
Wilton AN, Steward DJ, Zafiris K (1999) Microsatellite variation in the Australian dingo.
Journal of Heredity 90, 108-111.
Woodall PF, Pavlov P, Twyford KL (1996) Dingoes in Queensland, Australia: skull
dimensions and the indenity of wild canids. Wildlife Research 23, 581-587
Yuri T, Jernigan RW, Brumfield RT, Bhagabati NK, Braun MJ (2009) The effect of marker
choice on estimated levels of introgression across an avian (Pipridae: Manacus)
hybrid zone. Molecular Ecology 18, 4888-4903.
36
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
Data Accessibility
Data used in this research are lodged online in the Dryad data repository under
doi:10.5061/dryad.2rd32. This includes raw genotype and sample location data, 3Q test
results and an input file for the PCoA, and Structure input file and proportion of dingo
ancestry results.
Figure Legends.
Fig. 1. Locations of free-ranging dog specimens collected for this project, including
only those with a successfully determined genotype. The states and territory included
are Western Australia (WA, n = 2,284), Northern Territory (NT, n = 128), South
Australia (SA, n = 148), Queensland (QLD, n = 356), New South Wales (NSW, n =
95) and Victoria (VIC, n = 626). Specimens collected in the Australian Capital
Territory were classified as ‘NSW’.
Fig. 2. Principal coordinates analysis of specimens identified as being either ‘pure
dingo’ (category 1) or ‘domestic dog’ (7) from the average 3Q method results.
Clusters enclosed by rectangles were used for the a posteriori reference dingo
(coordinate 1 > 0.15) and domestic dog (coordinate 1 < -0.5) groups. The horizontal
dimension accounts for 66.2% of the variation, and an additional 8.1% is explained
by the vertical dimension. Inset map shows the geographic distribution of dingo
reference specimens (• = one specimen).
37
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
Fig. 3. Ordinary kriging of purity results and proportions of dingoes and hybrids by
state. State abbreviations are described in Fig 1. For kriging maps red-yellow patches
represent areas with a high level of dingo purity, blue regions contain individuals
with lower purity categories. Unsampled areas are shown in grey. (a) Average 3Q
method results with categories 1-6 interpolated. Major towns (>25,000 residents,
excluding capital cities) and their population sizes are shown for association with
hybridisation. Two smaller towns (+; Broome and Kununurra) are also shown
because they occur near higher levels of hybridisation than surrounding regions. (b)
Percentage of dingoes and hybrids in each Australian state, based on results from the
average 3Q method purity tests: dingoes are categories 1 and 2, hybrids categories 3-
6. (c) Ordinary kriging on the percentage of dingo ancestry from the clustering
method results. Results below 60% dingo ancestry are grouped as few specimens
were less than approximately half dingo. Towns referenced in the discussion are
labelled. (d) Percentage of dingoes and hybrids in each state as determined by the
clustering method. Dingoes are defined as specimens with ≥0.90 Q-assignment to the
dingo cluster.
Fig. 4. Comparison of expected component of dingo ancestry from simulated data
and the results from both the average 3Q and clustering methods. The x-axis values
are the simulated hybrid types 1-10 as classified in Table 2. Error bars show 95%
confidence intervals. (a) Results from the average 3Q method categories. Each box
for the ‘expected’ values shows the range of the average 3Q score for the appropriate
reference category. The expected values for crosses 2 and 7 cover two categories as
the expected value is in between them. (b) Clustering results showing the percentage
of dingo ancestry for each cross.
38
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
