Monitoring of genetic diversity in Taiwan conserved chickens assessed by pedigree and molecular data

Livestock Science 184 (2016) 85–91

Contents lists available at ScienceDirect

Livestock Science
journal homepage: www.elsevier.com/locate/livsci

Monitoring of genetic diversity in Taiwan conserved chickens
assessed by pedigree and molecular data
Manh-Hung Pham a,b,n, Xuan-Hoan Tran c, Cécile Berthouly-Salazar d,
Michèle Tixier-Boichard e, Chih-Feng Chen a,f, Yen-Pai Lee a
a

Department of Animal Science, National Chung-Hsing University, Taichung 40227, Taiwan
Faculty of Animal Science, Vietnam National University of Agriculture, Trau Quy Town, Gia Lam District, Ha Noi City, VietNam
c
Key Laboratory of Animal Cell Technology, National Institute of Animal Sciences, Tu Liem, Hanoi, VietNam
d
Institut de Recherche pour le Développement, UMR Diversité, Adaptation et Développement des Plantes (DIADE), Avenue Agropolis, BP 64501, 34394
Montpellier Cedex 5, France
e
INRA/AgroParisTech, UMR1313 Génétique Animale et Biologie Intégrative, Jouy-en-Josas, France
f
Research Center for Integrative and Evolutionary Galliformes Genomics (iEGG), National Chung-Hsing University, Taichung 40227, Taiwan
b

art ic l e i nf o

a b s t r a c t

Article history:

Received 9 August 2015
Received in revised form
27 December 2015
Accepted 29 December 2015

Local chicken breeds face high risks of extinction. A conservation program has been set up for eight
Taiwan conserved chicken populations (TCP). The research presented here aims at estimating effective
population size (Ne) and conservation priorities of TCP populations using pedigree and molecular data.
Genome diversity was assessed by genotyping 22 microsatellite markers in 45–50 animals per breed.
Results from the pedigree-based analysis showed that most Ne values ranged between 50 and 100 except
the Shek-Ki breed which exhibited the smallest value (46) so that most breeds could be considered as
safe from a conservation point of view. The change in inbreeding per generation varied between 0.7% to
1.9% depending on breeds. Ne values estimated from molecular-based analysis were generally lower than
those estimated from pedigree-based analysis, suggesting a loss of diversity between the onset of the
conservation program (from 1983 to 1995) and the start of pedigree recording in 2002. According to Ne
values, the TCP populations do not appear to be at a high risk, but mating plans by a rotation mating
system should be designed in order to limit the increase in inbreeding. Regarding the conservation
strategy within the TCP, the Shek-Ki and Hua-Tung breeds showed the highest priority for conservation
in terms of genetic risk status and contributions to total diversity across pedigree- and molecular-based
approaches. In conclusion, this study of TCP populations shows how different types of data can be
combined to define conservation priorities considering risk, diversity, or utility of local chicken breeds.
& 2016 Published by Elsevier B.V.

Keywords:
Conservation priorities
Effective population sizes
Inbreeding
Molecular data
Pedigree information

1. Introduction
Local chicken breeds play an important role in Taiwan due to the
traditional cuisine and culture. Local chicken breeds may carry
disease-resistant genes and show high abilities to adapt to alternative farming systems, such as organic, which will particularly
improve animal welfare and food safety (Fanatico et al., 2009; Pham
et al., 2012) as well as adaptation to harsh environmental conditions
(Tixier-Boichard et al., 2009). Phenotypic data and pedigree

n
Corresponding author at: Faculty of Animal Science, Vietnam National University of Agriculture, Trau Quy Town, Gia Lam District, Ha Noi City, VietNam.
E-mail addresses: (M.-H. Pham),
(X.-H. Tran),
(C. Berthouly-Salazar),
(M. Tixier-Boichard),
(C.-F. Chen), (Y.-P. Lee).

/>1871-1413/& 2016 Published by Elsevier B.V.

information have proven to be useful for characterization and
management of genetic diversity (Boichard et al., 1997; Tixier-Boichard et al., 2009; Lenstra et al., 2012). Unfortunately, phenotypic
data and pedigree records of local chickens are rarely documented
in reality (Tixier-Boichard et al., 2009). Therefore, molecular markers are used to monitor the loss of genetic diversity of populations
and set priorities for conservation (Boettcher et al., 2010).
FAO (2014) reported that 21.3 percent of chicken breeds in the
world were classified as being at risk of extinction, highlighting
the importance to assess genetic diversity and the current population status. This percentage might be higher than that because of
a large number of populations with an unknown status in developing countries. Basically, there are three strategies for setting
priorities in conservation such as the maximum-risk strategy, the
maximum-diversity strategy as well as the maximum-utility
strategy (Bennewitz et al., 2007).

The maximum-risk strategy is based on the numbers of


86

M.-H. Pham et al. / Livestock Science 184 (2016) 85–91

breeding animals (FAO, 2000), inbreeding rate (EAAP, 1998; Meuwissen, 2009) and recommended effective population size (Ne)
(Meuwissen, 2009). The Ne measures the number of breeding individuals in an idealized population in equilibrium that would
show a similar trend in inbreeding as the population under study,
it is one of the most pivotal parameters in both evolutionary
biology and conservation of genetic diversity (Waples and Do,
2010; Goyache et al., 2011; Leroy et al., 2013). This parameter is
considered as one of the major criteria for monitoring risk status
in livestock populations because it accounts for inbreeding and
loss of genetic diversity through random genetic drift (Falconer
and Mackay, 1996; Meuwissen, 2009). Leroy et al. (2013) showed
that estimates of effective population size varied according to the
within-breed genetic structure, for different species (i.e. cattle,
dog, horse and sheep).
The maximum-diversity strategy defines that a breed is selected for conservation when it contributes significantly to the
overall genetic diversity weighted by both the between- and
within-breed diversity. Breeds can be ranked according to their
contribution either to the actual or to the predicted future diversity (Bennewitz et al., 2007). For instance, Zanetti et al. (2010)
set up conservation priorities for five Italian local chicken breeds
undergoing in situ conservation using 20 microsatellites. In the
case of Vietnamese domestic chickens, Pham et al. (2013a) showed
that black H’mong, Lien Minh and Luong Phuong were ranked
with the highest priorities for conservation according to Caballero
and Toro (2002), and Petit et al. (1998) approaches, taking into

account within- and between-breeds components of diversity.
When possible, pedigree information and molecular data
should be combined for decision making of conservation priorities
(Zanetti et al., 2010). Pedigree-based and molecular-based estimates of genetic diversity may be more or less correlated depending on the pedigree completeness and the number of markers
(Toro et al., 2006) as illustrated in Iberian pigs with correlations
between pedigree inbreeding and marker homozygosity ranging
from 0.69 with 49 microsatellites (Toro et al., 2002) to 0.92 with
60 K SNPs (Silió et al., 2010). Furthermore, perfect correlations
between approaches cannot be reached because pedigree-based
estimates do not take into account Mendelian sampling, and it is
known that full-sibs would share between 45% and 55% of their
genes rather than exactly 50% in a traditional relationship matrix
(VanRaden and Tooker, 2007). An additional difference between
pedigree- and molecular-based analyses is the definition of the
founder population, which depends on the depth of pedigree for
pedigree-based analysis, the more complete the pedigree, the
more ancient the founder population (Falconer and Mackay, 1996).
Consequently, pedigree- and molecular-based analysis is using
different information, as pedigree-based analysis reflects only diversity due to relatively recent ancestry, depending on the population history (Toro et al., 2006; Engelsma et al., 2012).
The purposes of this study were (i) to assess genetic diversity
with pedigree-based estimates and molecular-based estimates for
eight populations kept under a conservation program in Taiwan
since 1982, and (ii) to monitor trends in genetic diversity and to
make recommendations for conservation strategy.

2. Material and methods
2.1. Data
Conservation of native chickens in Taiwan started from 1982,
when native chickens were collected around the islands and
conserved at National Chung-Hsing University (NCHU) experimental farm. Eight Taiwan conserved chicken populations (TCP: B

strain, L2 strain, Hsin-Yi, Hua-Tung, Ju-Chi, Nagoya, Quemoy and

Table 1
Pedigree information in the first generation of the eight populations when conservation program started.
Population

First generation

B strain (BS)
L2 strain (LS)
Hsin-Yi (HY)
Hua-Tung (HT)
Ju-Chi (JC)
Nagoya (NG)
Quemoy (KM)
Shek-Ki (KT)

Year

Sire

Dam

Nes

ΔF

1984
1984
1984

1990
1986
1989
1995
1989

6
5
7
2
12
9
1
6

20
28
15
4
38
22
4
19

18.5
17.0
19.1
5.3
36.5
25.5

3.2
18.2

2.71
2.95
2.62
9.38
1.37
1.96
15.63
2.74

Nes, effective population size based on number of sires and dams; and ΔF, rate of
hypothetical inbreeding (in percentage) for a population with such an effective
population size.

Shek-Ki) have been conserved at NCHU experimental farm since
then (Lee, 2006). The L2 and B strains were selected by NCHU from
the same Taiwan native chicken population (Lee, 2006). Both
strains were closed populations since their establishment in 1983,
while B strain was a male line and a L2 female line for crossing to
produce commercial meat-type chicken. Since then, they have
been selected for 24 and 26 generations, respectively, and have
been extensively used in research as well as in production (Chao
and Lee, 2001; Chen et al., 2007; Pham et al., 2013b). A small
number of parents were used to set up the first generation for
eight populations between 1984 and 1995 (Table 1). The management of conserved chicken populations followed a routine
procedure (Chao and Lee, 2001). Chicks were raised in floor pens
until 16 week of age, when they were transferred to individual
wire floored cages. Artificial insemination was individually used

for the female chickens with sire known and dam known. On the
average, the generation interval of TCP populations was one generation per year (Table 2). For the present study, pedigree information recorded between 2002 (starting with ancestors in
2001) and 2008 were used to estimate the effective population
size with different methods. The pedigree information included a
total of 4283 individuals and the numbers of founders at the onset
of pedigree recording are shown in Table 2. In addition, samples
from 383 individuals (i.e. 288 individuals from six TCP born in
2003 and 95 individuals from B and L2 strains born in 2008) were
genotyped and part of the data was previously published (Berthouly et al., 2008; Chang et al., 2012; Pham et al., 2013b). Briefly,
an average of 48 individuals per population was genotyped for 22
microsatellites among FAO (2011) recommended markers. These
Table 2
Number of founders at the onset of pedigree recording and average number of male
and female in the 2002–2008 periods for the eight populations.
Population

B strain
L2 strain
Hsin-Yi
Hua-Tung
Ju-Chi
Nagoya
Quemoy
Shek-Ki

Founders in 2001

2002–2008 generations

Sire

Dam

Nes

ΔF

Nm

Nf

Nes

ΔF

g2008

5
16
22
19
19
21
25
16

35
328
30
40

36
39
44
35

17.5
61.0
50.8
51.5
49.7
54.6
63.8
43.9

2.86
0.82
0.98
0.97
1.01
0.92
0.78
1.14

11.2
20.8
17.6
17.7
20.1
17.7
16.7

13.6

76.8
194.2
34.1
37.4
40.4
43.4
43.1
32.6

39.1
75.2
46.4
48.1
53.8
50.3
48.2
38.3

1.28
0.67
1.08
1.04
0.93
0.99
1.04
1.30

6

5
6
6
6
6
6
6

Nes, effective population size based on number of founder animals; ΔF, rate of
hypothetical inbreeding (in percentage) expected for a population with Nes; Nm,
average number of male; Nf, average number of female; Nes, effective population
size based on number of breeding animals; and ΔF, rate of hypothetical inbreeding
(in percentage) expected for a population with Nes across the 2002–2008 generations; g2008, number of generations known for the last generations.


M.-H. Pham et al. / Livestock Science 184 (2016) 85–91

samples were used to estimate the contemporary effective population size and contributions to diversity.
2.2. Demographic and Pedigree-based analysis
2.2.1. Effective number of founders
The effective number of founders (fe) is the number of equally
contributing founders, which would give the same amount of
genetic diversity that is present in the current population. This
was calculated as in Lacy (1989), and it is usually much smaller
than the actual number of founders in pedigree (animals with both
parents unknown) because of unequal contributions of founders to
the current population.
2.2.2. Effective number of ancestors
The effective number of ancestors (fa) was calculated as in
Boichard et al. (1997), explaining the complete genetic diversity of

a population. When compared with the effective number of
founders, it provides evidence of bottlenecks that occurred in
population in the past.
2.2.3. Effective number of founder genomes
The loss of genetic diversity would occur due to genetic drift in
a small population, even if founders would contribute equally to
this population (Ballou and Lacy, 1995). Therefore, the effective
number of founder genomes or founder genome equivalent (fge) is
defined as the number of equally contributing founders with no
loss of founder alleles that would give the same amount of genetic
diversity as is present in the reference population. The fge was
calculated as in Caballero and Toro (2000). It accounts for the loss
of genetic diversity that occurred in the population due to genetic
drift and bottlenecks.
2.2.4. Effective number of nonfounders
The effective number of nonfounders (Nenf) was calculated as
⎛ 1 1 ⎞−1
Nenf = ⎜ f – f ⎟ , where Nenf accounts for the contributions of
⎝ ge e ⎠
nonfounders and for loss of genetic diversity due to drift accumulated over nonfounder generations (Caballero and Toro, 2000).
2.2.5. Measure of the loss of genetic diversity
Measures of the loss of genetic diversity can be derived from fe,
fge, and Nenf. The amount of genetic diversity (GD) in the reference
population was computed according to Lacy (1995) as
1
GD = 1 − 2f . When expressed as 1 ÀGD, it measures the genetic
ge

diversity loss in the population since the founder generation, as a
1

result from both genetic drift and bottlenecks. GD* = 1 – 2f ,
e

*

1 À GD measures the loss of genetic diversity that occurred in the
population due to the unequal contributions of founders before
their contributions converged (Caballero and Toro, 2000). The
1
difference between GD* and GD is GD* − GD = 2N , which meaenf

sures the loss of diversity by genetic drift accumulated over nonfounder generations (Caballero and Toro, 2000).
2.2.6. Effective population size based on the mating plan
The estimation of effective population size based on number of
sires and dams (Nes) follows Wright’s (1931) model. This method
makes possible to predict Nes under several assumptions, including random mating, absence of selection and random variation
of progeny size across parents (Leroy et al., 2013). Computation
of Nes only requires the estimated numbers of breeding males
(Nm) and females (Nf) in the reference population using the

equation: Nes =

4Nm Nf
.
Nm +Nf

87

The increment of hypothetical inbreeding

(ΔF) is inversely proportional to the number of Nes: ΔF =

1
.
2Nes

2.2.7. Effective population size based on individual inbreeding rate
The effective population size based on the individual increase
in inbreeding (Nei) was computed for each population (Gutiérrez
et al., 2008). The individual increase in inbreeding is defined as
g −1
ΔFi = 1– i 1– Fi , where Fi is the inbreeding coefficient for each
individual i, and gi is the equivalent complete generations (Gutiérrez et al., 2009). The mean of the ΔFi values computed for the n
individuals belonging to a given population of individuals (ΔF ) can
be used to estimate Nei by the method of Cervantes et al. (2008)
1
for each of these reference populations as Nei = 2ΔF .
2.2.8. Effective population size based on individual coancestry rate
The realised effective population size (Nec) was calculated for
each population based on the individual increase in coancestry
rates (Cervantes et al., 2011). The increase in coancestry between
(gj + g k )

any pair of j and k can be computed as Δcjk = 1 − 2 1 − Cjk ,
where cjk is the inbreeding of an offspring from j and k, and gj and
gk are the number of equivalent complete generations for individual j and k, respectively. By averaging the increase in coancestry for all pairs of individuals in a reference subpopulation, we
can estimate an effective population size based on coancestries as
1
Nec = 2Δc . The confidence interval of the estimated values of Nei
and Nec can be computed using the variance of ΔFi and Δcjk from

the individuals in each population (Gutiérrez et al., 2008). The
genealogical information was analyzed using the program Endog
v4.8, a computer program for monitoring genetic variability of
populations using pedigree information (Gutiérrez and Goyache,
2005).
2.3. Degree of nonrandom mating
The degree of nonrandom mating (α) was measured by the
correlation of genes within individuals relative to the correlation
of genes taken at random from the population as in Caballero and
Toro (2000). This coefficient gives an indication of the deviation
from Hardy–Weinberg equilibrium expectations and it is related to
inbreeding and coancestry coefficient by (1 À F)¼(1 À f) Â (1 À α),
where F and f are the inbreeding and coancestry coefficients, respectively (Wright, 1969).
2.4. Molecular-based analysis
2.4.1. Microsatellites genotyping and polymorphism
The presence of null alleles was tested using FreeNA software
(Chapuis and Estoup, 2007) in which loci with estimated frequencies of null alleles above 0.2 are considered to be potentially
problematic for calculations. The null allele frequency estimated
for the 22 loci was lower than 0.2 (data not shown) so we assumed
that null alleles were absent and used for our analyses. The matrix
of Nei's DA genetic distances (Nei et al., 1983) was computed by
Populations package 1.2.32 (Langella, 1999).
2.4.2. Contemporary effective population size
Contemporary effective population size (NeLD) was computed
from genotypic data by a point estimation method using linkage
disequilibrium (Hill, 1981; Waples, 2006). This method was implemented in LDNe program (Waples and Do, 2008), which corrects for biases resulting from the presence of a wide range of
sample sizes and rare alleles, and was developed by Waples


88

M.-H. Pham et al. / Livestock Science 184 (2016) 85–91

(2006). The NeLD could be calculated for unlinked loci
1
as NeLD =
, where r is correlation among alleles and S is
1
(3 × ( r 2 –

S

the following formula: D¼ FST Â CB þ(1 À FST) Â CW.

))

sample size (Hill, 1981; Waples, 2006). All alleles with frequencies
less than the critical values (Pcrit) of 0.05 were excluded (Waples
and Do, 2008). A jackknife method was used to construct 95% CIs
of the estimates.
2.5. Contributions to diversity
The contribution of each breed to total genetic diversity was
computed by three approaches: a method based on molecular
coancestry (Caballero and Toro, 2002), a method based on allelic
richness (Petit et al., 1998) and a Weitzman approach modified by
Ollivier and Foulley (2005).
The method described by Caballero and Toro (2002) uses as a
criterion for the maintenance of the maximum overall Nei’s (1987)
gene diversity (GD) minimizing the average of molecular kinship
within subpopulations (fs), the average of molecular coancestry

within metapopulation (fm) and the average of Nei's minimum
genetic distance between subpopulations (Nei, 1987). Total gene
diversity (GDT) is GDT ¼1 À fm. Genetic diversity within subpopulations is GDW ¼1 À fs. Genetic diversity between subpopulations is GDB ¼ fs À fm. This approach estimates the genetic diversity
remaining when removing a breed. Therefore, a positive value will
indicate that higher diversity is obtained when a breed is not included in the dataset.
The method described by Petit et al. (1998), is using the rarefied
number of alleles per locus, and was applied to assess the contribution of each subpopulation to total allelic richness (CT) in
meta-population. The CT included the AR of within-subpopulation
diversity (CS) and its divergence from other subpopulations (CD)
and therefore taking into account private alleles. In contrast to the
method of Caballero and Toro, this one estimates a contribution to
genetic diversity. Therefore, positive value would indicate that
when population is included into the dataset it would increase
genetic diversity. Contributions of the breeds to diversity were
computed using Molkin 3.0, a computer program for genetic
analysis of populations using molecular coancestry information
(Gutiérrez et al., 2005).
The method described by Ollivier and Foulley (2005), contributions to between-breeds diversity (CB) were assessed by
using marginal loss of genetic diversity, based on DA genetic distances (Nei et al., 1983), following Weitzman approach (Weitzman,
1993) implemented in WEITZPRO (Derban et al., 2002). Withinbreed contributions to diversity (CW) and aggregate diversity (D)
were calculated as suggested by Ollivier and Foulley (2005). The D
index was obtained after weighting CB by FST and CW by 1 ÀFST as

3. Results and discussion
3.1. Pedigree-based analysis
3.1.1. Inbreeding and effective population size
The estimates of hypothetical inbreeding increment per generation are given in Table 2. The highest ΔF was approximately
1.3% found in Shek-Ki and B strain that means that in average 1.3%
of heterozygosity was lost per generation. Table 3 showed that the
highest increase in inbreeding rates was found for Ju-Chi, HuaTung and Shek-Ki populations at the values of 1.34%, 1.76% and

1.88%, respectively, which exceeded the maximum level of one
percent per generation as recommended by Meuwissen (2009).
Additional management such as mating decisions (avoidance of
mating between relatives by forming 20 mating pairs) and optimum contribution selection (Grundy et al., 1998) should be considered. Avoiding mating between close relatives can be easily
performed by a rotation mating system where each male ‘X’ is
mated to a group of sisters/half-sisters, and the son of this male ‘X’
will be mated to the daughters of another male ‘Y’ which has no
common parents with ‘X’. Another modern way is to calculate
coancestry coefficients and to design the mating plan in a way to
minimize the average coancestry of parents by the software. Family sizes should be kept homogenous, with always one son kept
per sire family and at least one daughter kept per dam. An increase
of generation interval for these breeds could also minimize inbreeding and thus a better conservation of their genetic diversity.
The estimation of Ne values highly varied between breeds and
according to methods. Averaged Nes between 2002–2008 values
ranged from 38.3 in Shek-Ki to 75.2 in L2 strain, and were always
lower than those estimated for 2001, except for B and L2 strains
(Table 2). All but the Shek-Ki breed (Nei: 45.9) showed Nei values
within the recommended minimum levels between 50 and 100 to
maintain genetic variation for a population in a long term according to Meuwissen (2009). The average coancestry rates, corresponding to the expected inbreeding rates in the next generation, varied from 0.77 in L2 strain to 3.14% in Shek-Ki (Table 3). The
Nec values were approximately equal to one-third of the Nei values,
as indicated by the Nec/Nei ratio, and ranged from 21.0 in Shek-Ki to
35.9 in L2 strain (Table 3; Fig. 1). The ratio between Nec and Nei
values was lower than one, which reflected that no substructure
(Wahlund effect) is found within eight populations (Cervantes
et al., 2011).
Correlations between Nes and other Ne measures (Nei, Nec) were
positive but not significant (Spearman's rank correlation ¼0.51,

Table 3
Estimates of effective population sizes for each population obtained by the pedigree and molecular data.

Population

B strain
L2 strain
Hsin-Yi
Hua-Tung
Ju-Chi
Nagoya
Quemoy
Shek-Ki

Pedigree informationa

Linkage disequilibriumb

n

ΔF

Nei 7 SE

Δc

Nec 7 SE

Nec
Nei

n

r

NeLD (95% CI)

595
1346
362
386
424
428
419
323

0.69
0.61
0.75
1.76
1.34
0.79
0.91
1.88

77.8 7 12.0
98.1 7 23.7
72.7 7 27.6
59.3 7 12.3
72.7 7 15.5
76.1 7 14.0
98.7 7 12.8
45.9 7 8.3

1.58
0.77
1.61
2.60
2.47
1.66
2.19
3.14

28.1 71.8
35.972.2
25.4 72.4
22.3 71.8
24.671.9
25.7 72.2
25.5 71.9
21.0 71.8

0.36
0.37
0.35
0.38
0.34
0.34
0.26
0.46

45
50

48
48
48
48
48
48

0.032
0.030
0.040
0.040
0.031
0.031
0.031
0.042

57.7
57.8
16.1
19.9
36.6
38.9
52.7
16.7

(27.4, 305.1)
(34.1, 131.7)
(12.2, 21.3)
(15.0, 26.9)
(23.1, 66.5)

(23.8, 75.4)
(34.4, 94.7)
(12.0, 23.5)

a
n, number of individuals; ΔF , the rates of inbreeding per generation (in percentage); Nei, effective population size based on individual increase in breeding; Δc , the rates
of coancestry per generation (in percentage); Nec, effective population size based on increase in coancestry; and ratio Nec .
Nei
b
n, number of individuals, r, the estimated correlation among alleles; and NeLD, contemporary effective population size
for each population.


M.-H. Pham et al. / Livestock Science 184 (2016) 85–91

Fig. 1. Effective population sizes for each population obtained by the pedigree and
molecular data. Nei, effective population size based on individual increase in
breeding with its standard error bars; Nec, effective population size based on increase in coancestry with its standard error bars; and NeLD, contemporary effective
population size for each population.

0.57, P4 0.05).
3.1.2. Genetic contributions
For the period between 2002 and 2008, the estimated values of
the fe, fa, fge, Nenf and Na50 in eight populations are given in Table 4.
They were always higher than the real number of founders recorded in the early generations between 1983 and 1995. In ShekKi, the estimated values were consistently lower than in the remaining seven populations, except for Nenf, which was lower in
Hua-Tung. The fe, fa and fge values for Shek-Ki were 21, 20 and 16,
respectively. The numbers of ancestors needed to explain 50% of
the gene pool ranged from 7 in Shek-Ki to 16 in L2 strain. These
results indicate that Shek-Ki has a narrow genetic base and contributes less to the within-breed component of genetic diversity
than the remaining populations. The effective number of nonfounders was greatly higher than effective number of founder

genomes, which indicates that loss of genetic diversity due to drift
was accumulated over nonfounder generations.
The loss of overall genetic diversity observed in Hua-Tung and
Shek-Ki was 5.2% and 6.3%, respectively. The first loss ranged from
2.6% in Hua-Tung to 3.1% in Shek-Ki due to bottlenecks and genetic
drift (Fig. 2). The second loss accounted for unequal founder
contribution and ranged from 1.7% in Hua-Tung to 2.4% in Shek-Ki.
The remaining loss was due to genetic drift only.
3.2. Degree of nonrandom mating
The average inbreeding coefficients were smaller than average
Table 4
Genetic contributions of eight Taiwan conserved chicken populations.
Population

n

fe

fa

fge

Nenf

Na50

F

f

F2008 7 SE

B strain
L2 strain
Hsin-Yi
Hua-Tung
Ju-Chi
Nagoya
Quemoy
Shek-Ki

595
1346
362
386
424
428
419
323

44
87
41
29
28
38
31
21

21

45
38
27
27
31
26
20

31.6
65.5
31.1
19.3
20.3
30.1
22.8
16.0

112.1
265.0
128.8
57.7
73.8
144.8
86.2
67.2

8
16
14
10

10
11
10
7

0.011
0.006
0.010
0.019
0.013
0.008
0.009
0.020

0.017
0.007
0.017
0.026
0.025
0.018
0.022
0.031

0.0517 0.006
0.025 7 0.003
0.048 7 0.015
0.065 7 0.011
0.038 7 0.004
0.0217 0.003
0.0357 0.004

0.0727 0.013

n, number of individuals; fe, estimates of effective number of founders; fa, effective
number of ancestors; fge, effective number of founder genomes; Nenf, effective
number of nonfounders; Na50, number of ancestors explaining 50% of the gene
pool; F, inbreeding coefficient for each population in the 2002–2008 periods; f,
coancestry coefficient for each population in the 2002–2008 periods; and F2008,
inbreeding coefficient of the last generation in 2008; SE, standard error.

89

Fig. 2. Genetic diversity (GD) loss in eight populations in the period between 2002
and 2008. GD loss due to bottlenecks and genetic drift (1 À GD); due to unequal
founder contribution (1 À GD*); and genetic drift only (GD* À GD).

coancestries in all eight populations (Table 4). Therefore, the degree of nonrandom mating has been negative and mating of highly
average-related individuals was successfully avoided. The average
inbreeding rates were moderate and ranged from 2.1% in Nagoya
to 7.2% in Shek-Ki for the last generation (Table 4). However, these
values do not take into account the fact that the founders in 2001
could have been already inbred because each population was
closed since several generations. In practice, chickens showing a
low individual inbreeding should be chosen for mating in order to
maintain the recommended effective population sizes and to
prevent the risk of extinction in small populations.
3.3. Molecular-based analysis
3.3.1. Contemporary effective population size
The estimates of NeLD were lower than the ones of Nei given by
the pedigree-based analysis and that of Nes values (Tables 2 and 3;
Fig. 1). The correlations between NeLD and Nes was very low

(Spearman’s rank correlation ¼0.24); but it was significantly positive between NeLD, Nei and Nec (Spearman’s rank
correlation ¼0.85, 0.83 and 0.83, respectively, P o0.05). The sample size used for genotyping was included in the CIs of NeLD. It is
possible that this underestimation is due to the small number of
molecular markers used (Waples, 2006; Engelsma et al., 2012).
Engelsma et al. (2010) showed that marker density was important
in order to assess genetic diversity with heterozygosity estimates
but was not so much important when an IBD method was used to
assess diversity.
The difference between molecular-based estimate of Ne and
pedigree-based estimates of Ne could also occur if a loss of diversity took place between the initial founder generation and the
first generation included in the pedigree. Pedigree-based estimates are as good as the pedigree data, which, in the present case,
represents a quite recent part of the history of populations. According to Toro et al. (2006), some important loci might be highly
differentiated because selective forces are strong enough at such
loci to overcome the effect of low effective size. In order to
maintain genetic diversity in a gene bank of Holstein cattle, Engelsma et al. (2012) concluded that high density SNP-based diversity provided a more detailed knowledge of diversity at the
scale of chromosomal regions than pedigree-based estimate of
diversity, which remains global.
3.3.2. Contributions to diversity
The conservation priorities were mostly consistent across three
approaches we used. Contribution to global diversity (GDG) was
significantly negatively correlated with total allelic diversity (CT,
Spearman’s rank correlation¼ À 0.83, P o0.05) and with aggregate
diversity (D, Spearman's rank correlation¼ À 0.93, P o0.001). The


90

M.-H. Pham et al. / Livestock Science 184 (2016) 85–91

Table 5

Loss or gain of genetic diversity, contributions to allelic richness and aggregate diversity for each of the eight Taiwan conserved chicken breeds.
Population

B strain
L2 strain
Hsin-Yi
Hua-Tung
Ju-Chi
Nagoya
Quemoy
Shek-Ki

Loss/gain of diversity

Contribution to richness

Weitzman's diversity

GDW

GDB

GDG

CS

CD

CT

CW

CB

D

0.20
0.96
À 0.68
À 2.64
1.33
1.56
À 0.71
0.00

À 0.41
À 1.39
1.33
1.57
À 0.13
À 3.32
0.54
À 2.47

À 0.21
À 0.44
0.65
À 1.07
1.21
À 1.76

À 0.18
À 2.47

0.32
1.27
0.72
0.39
À 0.42
À 0.64
À 0.24
À 1.40

0.47
3.00
1.32
2.89
À 1.21
3.37
2.58
6.73

0.80
4.27
2.04
3.28
À 1.64
2.73
2.34
5.33

À 0.42
À 1.27
0.89
3.57
À 1.76
À 2.04
0.98
0.04

6.55
20.36
4.91
9.27
2.55
31.82
11.09
30.00

1.68
4.63
2.15
5.39
À 0.47
7.02
4.09
8.29

GDW, change in within-population genetic diversity after removing population i; GDB, change in between-population genetic diversity after removing population i; GDG,
change in global diversity after removing population i; CS, a contribution to within-population genetic diversity; CD, a between-population genetic diversity; CT, Total
diversity; CW, contribution to within-population diversity; CB, contribution to between-population diversity; and D, aggregate diversity (in percentage).

D index was positively correlated with the CT (Spearman's rank
correlation¼0.91, P o0.01). Negative correlation is expected with
GDG since it expresses a loss why other methods express a gain.
The results showed that Shek-Ki had the highest contributions
(GDG ¼ À 2.47%, CT ¼5.33% and D¼8.29%) to overall genetic diversity (Table 5). Nagoya ranked the second contributions to diversity (GDG ¼ À1.76% and D¼7.02%), and the fourth of CT ¼ 2.73%
for conservation. Hua-Tung contributed the third rank (GDG
¼ À 1.07%, CT ¼3.28% and D ¼5.39%), but exhibited the highest
GDW for conservation, whereas the traditional Ju-Chi breed had
the smallest values of diversity (GDG ¼1.21%, CT ¼ À 1.64% and
D¼ À 0.47%) for conservation priorities. Overall, the Shek-Ki, HuaTung and Nagoya had high conservation priorities in terms of
genetic diversity. However, Shek-Ki and Nagoya exhibited very low
within-breed diversity and it was highly differentiated from other
breeds and appeared as obvious conservation priorities. When
analysed for six TCP (i.e. not including B and L2 strains), the HuaTung contributed the highest to the aggregate diversity (Berthouly
et al., 2008). Thus, using a different set of breeds seems to change
the relative contribution of each breed, as observed in 24 Vietnamese domestic chicken populations (Pham et al., 2013a). In fact,
the main contribution of the Hua-Tung was based on its gene diversity and thus this breed had a high contribution to the gene
pool.
3.4. Potential conservation
We have analyzed genetic diversity for eight Taiwan conserved
chicken breeds based on both pedigree and molecular data. The
smallest values of effective population size based on pedigree information were observed in Shek-Ki and Hua-Tung breeds. Thus,
the Shek-Ki and Hua-Tung breeds showed high priorities for
conservation in terms of genetic risk status and contributions to
diversity. In spite of a very small number of founders in the first
generation, the Hua-Tung breed exhibited the highest contribution
to within-breed diversity. It is likely that the initial founders were
very diverse because of the breed history: farmers kept this breed
together with others in the backyard, so that Hua-Tung might have

been crossed with imported larger game birds from Southeast Asia
(Lee, 2006; Berthouly et al., 2008), which have bright black feather
and large body size. Quemoy exhibited also a relatively high NeLD
and Nei in spite of a very narrow base population. Contrasting with
the history of the Hua-Tung breed, the Quemoy breed is expected
to be a true native since it was essentially isolated from outside the
world between 1949 and around 1990. Although chickens were
also kept in the backyard, the original owners intentionally kept
them as pure as possible (Chia-Juing Won, personal communication). However, the Quemoy conservation program is the most
recent of all populations, since it started in 1995. Thus, it is likely

that effects of genetic drift have been less important in this breed
than in others. Therefore, the effective numbers of founder genomes of Hua-Tung and Quemoy showed high values, for different
reasons.
Considering the maximum-utility approach is also needed before making recommendations for conservation. Conservation
priorities should include the market demand, survivability and
productivity related to specific genes under existing management
conditions, scocio-economics and the needs for research and development (Toro et al., 2006; Pham et al., 2013a). In this respect,
Shek-Ki has been selected as a sire line to be distributed to farmers
producing the Three Yellow breed for meat consumption in Hong
Kong (Berthouly et al., 2008). This breed showed the highest body
weight at 16 weeks of age (Chang et al., 2012). Selection for a
specific trait has resulted in a reduction of genetic variation (Toro
et al., 2011). Chang et al. (2012) showed that Hsin-Yi, Shek-Ki and
Hua-Tung males exhibited a better heat tolerance due to a lower
panting rate than observed in others. In addition, Hsin-Yi responded to the highest antibody levels from Infectious Bursal
Disease (IBD) vaccine. The Quemoy exhibited high antibody response to low pathogenic avian influenza H6N1 virus, Newcastle
Disease and IBD vaccines.

4. Conclusions

The Shek-Ki breed is of high priority for conservation considering risk and utility, the Hua-Tung breed is of high priority
considering risk and diversity, and the Quemoy breed is of high
priority considering diversity and utility. Then, the Nagoya breed
may be of high priority considering between-breeds diversity, and
the Hsin-Yi, B and L2 lines could be considered for conservation on
the basis of utility only. Thus, this study of TCP populations shows
how different types of data can be combined to define conservation priorities considering risk, diversity, or utility of local breeds.

Conflict of interest statement
No actual or potential conflict of interest in relation to this article exists.

Acknowledgments
We sincerely appreciate the chicken caretaker at NCHU experimental farm and graduate students for their assistance in
pedigree recording. We also would like to thank two anonymous


M.-H. Pham et al. / Livestock Science 184 (2016) 85–91

reviewers to revise the manuscript with several improvements.

References
Ballou, J.D., Lacy, R.C., 1995. Identifying genetically important individuals for
management of genetic variation in pedigreed populations. In: Ballou, J.D.,
Gilpin, M., Foose, T.J. (Eds.), Population Management for Survival and Recovery:
Analytical Methods and Strategies in Small Population Management. Columbia
University Press, New York, NY, pp. 76–111.
Bennewitz, J., Eding, H., Ruane, J., Simianer, H., 2007. Selection of breeds for conservation. In: Oldenbrook, K. (Eds.), Utilization and Conservation of Farm Animal Genetic Resources, Wageningen Academic Publishers, Wageningen, The
Netherlands, pp. 131–146.
Berthouly, C., Bed’Hom, B., Tixier-Boichard, M., Chen, C.F., Lee, Y.P., Laloë, D., Legros,
H., Verrier, E., Rognon, X., 2008. Using molecular markers and multivariate

methods to study the genetic diversity of local European and Asian chicken
breeds. Anim. Genet. 39, 121–129.
Boettcher, P.J., Tixier-Boichard, M., Toro, M.A., Simianer, H., Eding, H., Gandini, G.,
Joost, S., Garcia, D., Colli, L., Ajmone-Marsan, P., Consortium, G., 2010. Objectives, criteria and methods for using molecular genetic data in priority setting
for conservation of animal genetic resources. Anim. Genet. 41 (Suppl. 1),
S64–S77.
Boichard, D., Maignel, L., Verrier, E., 1997. The value of using probabilities of gene
origin to measure genetic variability in a population. Genet. Sel. Evol. 29, 5–23.
Caballero, A., Toro, M.A., 2000. Interrelations between effective population size and
other tools for management of conserved populations. Genet. Res. 75, 331–343.
Caballero, A., Toro, M.A., 2002. Analysis of genetic diversity for the management of
conserved subdivided populations. Conserv. Genet. 3, 289–299.
Cervantes, I., Goyache, F., Molina, A., Valera, M., Gutiérrez, J.P., 2011. Estimation of
effective population size from the rate of coancestry in pedigreed populations.
J. Anim. Breed. Genet. 128, 56–63.
Cervantes, I., Goyache, F., Molina, A., Valera, M., Gutiérrez, J.P., 2008. Application of
individual increase in inbreeding to estimate effective sizes from real pedigrees.
J. Anim. Breed. Genet. 125, 301–310.
Chang, C.S., Chen, C.F., Berthouly-Salazar, C., Chazara, O., Lee, Y.P., Chang, C.M.,
Chang, K.H., Bed’Hom, B., Tixier-Boichard, M., 2012. A global analysis of molecular markers and phenotypic traits in local chicken breeds in Taiwan. Anim.
Genet. 43, 172–182.
Chao, C.H., Lee, Y.P., 2001. Relationship between reproductive performance and
immunity in Taiwan country chickens. Poult. Sci. 80, 535–540.
Chapuis, M.P., Estoup, A., 2007. Microsatellite null alleles and estimation of population differentiation. Mol. Biol. Evol. 24, 621–631.
Chen, C.F., Shiue, Y.L., Yen, C.J., Tang, P.C., Chang, H.C., Lee, Y.P., 2007. Laying traits
and underlying transcripts, expressed in the hypothalamus and pituitary gland
that were associated with egg production variability in chickens. Theriogenology 68, 1305–1315.
Derban, S., Foulley, J.-L., Ollivier, L., 2002. WEITZPRO: Software for Analysing Genetic Diversity. INRA, Paris.
EAAP, 1998. Assessment of the degree of endangerment of livestock breeds.
Working group on Animal Genetic Resources, In: Proceedings of the 49th Annual Meeting of European Association for Animal Production, Warsaw, Poland.

EAAP Publication, Wageningen Academic Publishers, Wageningen, The
Netherlands.
Engelsma, K.A., Calus., M.P.L., Bijma, P., Windig, J.J., 2010. Estimating genetic diversity across the neutral genome with the use of dense marker maps. Genet.
Sel. Evol. 42, 12.
Engelsma, K.A., Veerkamp, R.F., Calus, M.P.L., Bijma, P., Windig, J.J., 2012. Pedigreeand marker-based methods in the estimation of genetic diversity in small
groups of Holstein cattle. J. Anim. Breed. Genet. 129, 195–205.
Falconer, D.S., Mackay, T.F.C., 1996. Introduction to Quantitative Genetics. Longman
Group, Essex, UK.
Fanatico, A.C., Owens, C.M., Emmert, J.L., 2009. Organic poultry production in the
United States: Broilers. J. Appl. Poult. Res. 18, 355–366.
FAO, 2000. World Watch List for Domestic Animal Diversity, Scherf, B.D. (Eds.).
Third Edition, Rome.
FAO, 2011. Molecular Genetic Characterization of Animal Genetic Resources. FAO
Animal Production and Health Guidelines, Rome (No. 9).
FAO, 2014. Status and trends of animal genetic resources. Commission on genetic
resources for food and agriculture. Eighth Session. Rome, 26–28 November
2014.
Goyache, F., Álvarez, I., Fernández, I., Pérez-Pardal, L., Royo, L.J., Lorenzo, L., 2011.
Usefulness of molecular-based methods for estimating effective population size
in livestock assessed using data from the endangered black-coated Asturcón
pony. J. Anim. Sci. 89, 1251–1259.
Grundy, B., Villanueva, B., Woolliams, J.A., 1998. Dynamic selection procedures for
constrained inbreeding and their consequences for pedigree development.
Genet. Res. 72, 159–168.
Gutiérrez, J.P., Cervantes, I., Goyache, F., 2009. Improving the estimation of realized
effective population sizes in farm animals. J. Anim. Breed. Genet. 126, 327–332.
Gutiérrez, J.P., Cervantes, I., Molina, A., Valera, M., Goyache, F., 2008. Individual

91

increase in inbreeding allows estimating realised effective sizes from pedigrees.
Genet. Sel. Evol. 40, 359–378.
Gutiérrez, J.P., Goyache, F., 2005. A note on ENDOG: a computer program for analysing pedigree information. J. Anim. Breed. Genet. 122, 172–176.
Gutiérrez, J.P., Royo, L.J., Álvarez, I., Goyache, F., 2005. MOLKIN v2.0: a computer
program for genetic analysis of populations using molecular coancestry information. J. Hered. 96, 718–721.
Hill, W.G., 1981. Estimation of effective population size from data on linkage disequilibrium. Genet. Res. 38, 209–216.
Lacy, R.C., 1989. Analysis of founder representation in pedigrees: founder equivalent and founder genome equivalents. Zoo Biol. 8, 111–123.
Lacy, R.C., 1995. Clarification of genetic terms and their use in the management of
captive populations. Zoo Biol. 14, 565–578.
Langella, O., 1999. POPULATIONS 1.2.32. Copyright© by CNRS UPR9034, France.
Lee, Y.P., 2006. Taiwan country chicken: a slow growth breed for eating quality.
In: Liao, C.W., Shih, B.L., Lee, M.L., Hsu A.L. and Cheng Y.S. (Eds.), Symposium 7–
10 November 2006, Scientific Cooperation in Agriculture between Council of
Agriculture, Taiwan, R.O.C. and Institut National de Recherche Agronomique,
France, Technical Bulletin of Livestock Research Institute 103, pp. 121–132.
Lenstra, J.A., Groeneveld, L.F., Eding, H., Kantanen, J., Williams, J.L., Taberlet, P., Nicolazzi, E.L., Sölkner, J., Simianer, H., Ciani, E., Garcia, J.F., Bruford, M.W., Ajmone-Marsan, P., Weigend, S., 2012. Molecular tools and analytical approaches
for the characterization of farm animal genetic diversity. Anim. Genet. 43,
483–502.
Leroy, G., Mary-Huard, T., Verrier, E., Danvy, S., Charvolin, E., Danchin-Burge, C.,
2013. Methods to estimate effective population size using pedigree data: examples in dog, sheep, cattle and horse. Genet. Sel. Evol. 45, 1.
Meuwissen, T.H.E., 2009. Towards consensus on how to measure neutral genetic
diversity? J. Anim. Breed. Genet. 126, 333–334.
Nei, M., 1987. Molecular Evolutionary Genetics. Columbia Univ. Press, New York,
NY, pp. 1–512.
Nei, M., Tajima, F., Tateno, Y., 1983. Accuracy of estimated phylogenetic trees from
molecular data. J. Mol. Evol. 19, 153–170.
Ollivier, L., Foulley, J.L., 2005. Aggregate diversity: new approach combining withinand between-breed genetic diversity. Livest. Prod. Sci. 95, 247–254.
Petit, R.J., Mousadik, A.E., Pons, O., 1998. Identifying populations for conservation
on the basis of genetic markers. Conserv. Biol. 12, 844–855.
Pham, M.H., Berthouly-Salazar, C., Tran, X.H., Chang, W.H., Crooijmans, R.P.M.A., Lin,

D.Y., Hoang, V.T., Y.P., L., Tixier-Boichard, M., Chen, C.F., 2013a. Genetic diversity
of Vietnamese domestic chicken populations as decision-making support for
conservation strategies. Anim. Genet. 44, 509–521.
Pham, M.H., Chang, W.H., Berthouly-Salazar, C., Lin, D.Y., Yungrahang, S., Wang, C.C.,
Lee, Y.P., Tixier-Boichard, M., Chen, C.F., 2013b. Genetic characterization of
Taiwan commercial native chickens ascertained by microsatellite markers. J.
Poult. Sci., 290–299.
Pham, M.H., Tran, X.H., Lee, Y.P., Lin, D.Y., Pham, D.L., Hoang, V.T., Tixier-Boichard,
M., Chen, C.F., 2012. Genetic diversity of the major histocompatibility complex
region in Vietnamese local chickens using the LEI0258 microsatellite marker. J.
Anim. Sci. Technol. 34, 11–19.
Silió, L., Fernández, A., Mercadé, A., Martin-Palomino, P., López, M.A., Rodrigáñez, J.,
Ovilo, C., 2010. Measuring inbreeding in a closed pig strain from high-density
SNPs genotypes. In: Proceedings of the 9th World Congress Genetics Applied
Livestock Production Congress, Leipzig, Germany, 1–6 August 2010.
Tixier-Boichard, M., Bordas, A., Rognon, X., 2009. Characterisation and monitoring
of poultry genetic resources. Worlds Poult. Sci. J. 65, 272–285.
Toro, M.A., Barragán, C., Óvilo, C., Rodrigañez, J., Rodriguez, C., Silió, L., 2002. Estimation of coancestry in Iberian pigs using molecular markers. Conserv. Genet.
3, 309–320.
Toro, M.A., Fernández, J., Caballero, A., 2006. Scientific basis for policies in conservation of farm animal genetic resources. The 8th World Congress on Genetics Applied to Livestock Production, Belo Horizonte, MG, Brasil, August 13–
18, 2006.
Toro, M.A., Meuwissen, T.H.E., Fernández, A., Shaat, I., Mäki-Tanila, A., 2011. Assessing the genetic diversity in small farm animal populations. Animal 5,
1669–1683.
VanRaden, P.M., Tooker, M.E., 2007. Methods to explain genomic estimates of
breeding value. J. Dairy Sci. 90 (Suppl. 1), S374.
Waples, R.S., 2006. A bias correction for estimates of effective population size based
on linkage disequilibrium at unlinked gene loci. Conserv. Genet. 7, 167–184.
Waples, R.S., Do, C., 2008. LDNe: A program for estimating effective population size
from data on linkage disequilibrium. Mol. Ecol. Notes 8, 753–756.
Waples, R.S., Do, C., 2010. Linkage disequilibrium estimates of contemporary Ne

using highly variable genetic markers: a largely untapped resource for applied
conservation and evolution. Evol. Appl. 3, 244–262.
Weitzman, M.L., 1993. What to preserve? An application of diversity theory to
crane conservation. Q. J. Econ. 108, 157–183.
Wright, S., 1931. Evolution in Mendelian populations. Genetics 16, 97–159.
Wright, S., 1969. Evolution and the Genetics of Populations, Vol. 2: Theory of Gene
Frequencies. University of Chicago Press, Chicago, IL.
Zanetti, E., De Marchi, M., Dalvit, C., Cassandro, M., 2010. Genetic characterization of
local Italian breeds of chickens undergoing in situ conservation. Poult. Sci. 89,
420–427.