Two quantitative trait loci on Sus scrofa chromosomes 1 and 7
affecting the number of vertebrae1
S. Mikawa*2, T. Hayashi*, M. Nii†, S. Shimanuki‡, T. Morozumi‡, and T. Awata*
*Genome Research Department, National Institute of Agrobiological Sciences, Tsukuba, Ibaraki 305-0901,
Japan; †Livestock Research Institute, Tokushima Agriculture, Forestry and Fisheries Technology Center,
Anan, Tokushima 774-0047, Japan; and ‡STAFF Institute, Tsukuba, Ibaraki 305-0854, Japan
ABSTRACT: The objective of the research was to
identify QTL affecting the number of vertebrae in
swine, one of the major determining factors of growth
and body composition. Previously, we reported a QTL
for the number of vertebrae located on SSC1qter (terminal band of the q arm of SSC 1) in an F2 family produced
by crossing a Göttingen miniature male with two Meishan females. Eight other swine families were subsequently produced by crosses between different breeds of
European, Asian, and miniature pigs. In these families,
the QTL on SSC1qter for the number of vertebrae was
detected. Unlike the Asian alleles, all European alleles
in this study had the effect of increasing the number
of vertebrae by 0.44 to 0.69 and acted additively without
dominance. The Göttingen miniature sire, for which we
previously reported a smaller additive effect, seemed
to be heterozygous at the QTL. In the present study,
another QTL was found for the number of vertebrae on
SSC7. This QTL was not fixed in the European pigs
used as parents in our experimental families, and some
of the European alleles increased the number of vertebrae. A half-sib analysis confirmed that this QTL was
segregating in a commercial Large White population.
Analysis in an F2 family in which the parental pigs were
fixed for alternative alleles revealed that the effects of
the QTL on SSC1 and on SSC7 were additive and similar in size. The two QTL acted independently without
epistatic effects and explained an increase of more than
two vertebrae.
Key Words: Pigs, Quantitative Trait Loci, Swine, Vertebrae
2005 American Society of Animal Science. All rights reserved.
Introduction
For many years, it has been known that the number
of vertebrae in pigs varies. The number of cervical vertebrae is fixed at seven, as in other mammals, but the
thoracic and lumbar vertebrae vary in number (King
et al., 1960). In the King et al. (1960) report, the number
of thoracic vertebrae ranged from 14 to 16, and the
number of lumbar ranged from five to seven, in commercial pigs. The number of vertebrae affects carcass
length, and an increase of approximately 15 mm for
each additional vertebra can be expected in a carcass
approximately 800 mm long. Today, the total number
of thoracic and lumbar vertebrae in European breeds
ranges from 21 to 23. Wild boars, which are ancestors of
1
This work was supported by the DNA Marker Project of the Ministry of Agriculture, Forestry, and Fisheries of Japan and by a Grantin-Aid from the Japan Racing Association.
2
Correspondence: Ikenodai 2 (phone: 81-29-838-8627; fax: 81-29838-8627; e-mail: [email protected]).
Received November 3, 2004.
Accepted June 16, 2005.
J. Anim. Sci. 2005. 83:2247–2254
pigs, have a uniform number of 19 thoracic and lumbar
vertebrae. European breeds have been improved for
years, as body size has been enlarged to increase meat
production. We can assume that, in the process, the
number of vertebrae has increased as well.
In our previous study, we reported a QTL analysis for
various traits in an F2 family produced at the National
Institute of Animal Industry (NIAI) of Japan (referred
to here as the NIAI family) by crossing a Göttingen
miniature (Porter, 2002) male and two Meishan females, and detected QTL for various traits (Wada et
al., 2000). Among them, we detected a QTL on SSC1qter
(terminal band of the q arm of SSC1), which affected
the number of vertebrae. Interestingly, alleles of the
Göttingen miniature sire had the effect of increasing
the number of vertebrae, despite the smaller body size
of this pig. For further investigation of QTL affecting
economic traits, we produced eight other F2 families
by crossing different breeds of Asian, European, and
miniature pigs. In the present study, we used these
families to detect novel QTL affecting the number of
vertebrae, as well as that on SSC1qter, and analyzed
the variation of these QTL in commercial pigs.
2247
2248
Mikawa et al.
Table 1. Summary of swine experimental families
Breed of parentsa
No. of F1 progeny
Family
Grand dam
Grand sire
Dam
Sire
No. of F2
progeny
MGb
MLc
LWbb
JW
BC
MW
LM
JDd
WWbb
Meishan (2)
Meishan (2)
Landrace (2)
Jinhua (1)
Berkshire (1)
Meishan (1)
Landrace (1)
Jinhua (5)
Large White (3)
Göttingen Mini (1)
Landrace (1)
Wild Boar (1)
Large White (1)
Clawn Mini (1)
Large White (1)
Meishan (1)
Duroc (1)
Wild Boar (1)
18
7
5
4
6
4
7
21
7
2
1
2
1
2
1
1
6
3
154
138
127
122
119
84
116
528
353
a
Numbers of grand dams and grand sires are presented in parentheses.
An F1 sire derived from each grand dam was mated to full-sib F1 dams.
An F1 sire was mated to four full-sib F1 dams and three half-sib F1 dams.
d
F1 sires were derived from three of five grand dams, and they were mated to half-sib F1 dams.
b
c
Materials and Methods
Swine Populations and Traits
The MG family was produced by crossing a Göttingen
miniature male with two Meishan females. This MG
family, consisting of the NIAI family, was used for the
construction of a linkage map by Mikawa et al. (1999)
and for the QTL analysis by Wada et al. (2000). Additional F2 animals were produced later. Eight novel
three-generation families were produced by crossing
a Landrace male with two Meishan females (ML), a
Japanese wild boar with two Landrace females (LWb),
a Large White male with a Jinhua female (JW), a Clawn
miniature (Nakanishi, 1981; Kamimura et al., 1996)
male with a Berkshire female (BC), a Large White male
with a Meishan female (MW), a Meishan male with a
Landrace female (LM), a Duroc male with five Jinhua
females (JD), and a Japanese wild boar with three
Large White females (WWb; Nii et al., 2005). The numbers of F1 and F2 individuals in each family are listed
in Table 1. The JW and JD families were bred under
specific-pathogen-free conditions, and the others were
bred conventionally. All F2 animals were weaned at 28
d of age, and males were castrated. In the ML, JW,
MW, LM, and JD families, F2 animals were slaughtered
at approximately 70 kg of live weight. Those of the BC
family were slaughtered at 80 kg, and those of the LWb
family were slaughtered at 90 kg. In the MG family,
male pigs were slaughtered at 91 d of age, and females
were slaughtered after the first estrous cycle. After
slaughter, the numbers of thoracic and lumbar vertebrae were scored.
culated by a permutation test (Churchill and Doerge,
1994) with 5,000 repetitions. We used 293, 138, and
152 microsatellite markers in the USDA linkage map
(Rohrer et al., 1996) for the MG, ML, and LWb families,
respectively. In these analyses, map positions of the
USDA linkage map were used. For each of the JW, BC,
MW, LM, JD, and WWb families, the existence of QTL
was investigated by analyzing, with interval mapping,
each of SSC1 and SSC7, on which QTL for the number
of vertebrae were detected in genome scanning in the
MG, ML, or LWb families. The model used was the
same as described previously. We used 11, 15, 16, 6, 8,
and 21 markers on SSC1, and 8, 12, 8, 6, 9, and 9
markers on SSC7 in the JW, BC, MW, LM, JD, and
WWb families, respectively. The USDA linkage map
positions were used for the analyses. The chromosomewide significance threshold was calculated by a permutation test (Churchill and Doerge, 1994) for each of
SSC1 and SSC7, with 5,000 repetitions. We also estimated the genome-wide significance threshold from the
chromosome-wide significance threshold following the
method of de Koning et al. (1999):
Pgenomewise = 1 − (1 − Pchromosomewise)1/r.
The contributions (r) of SSC1 and SSC7 to the total
length of autosomal chromosome were 0.0667 and
0.0726, respectively, calculated using the report of Rohrer et al. (1996). Genomewise P = 0.01 corresponds to
SSC1 P = 6.7 × 10−4 and SSC7 P = 7.3 × 10−4. Genomewise P = 0.05 corresponds to SSC1 P = 3.4 ×
10−3 and SSC7 P = 3.7 × 10−3.
QTL Analysis for the Number of Vertebrae
Evaluation of Heterozygosity of QTL on SSC1
of the Göttingen Miniature Sire in the MG Family
A QTL scan of the genome was performed for the
number of vertebrae in each of the MG, ML, and LWb
families. An interval mapping approach developed by
Haley et al. (1994) was used. The model used assumed
that the founder breeds were fixed for alternative alleles. The genome-wide significance threshold was cal-
To scrutinize the heterozygosity of the QTL on SSC1
in the MG family, we inferred the inheritance of each
homologous allele at the QTL in the Göttingen miniature sire to F1 and F2 individuals by using the microsatellite marker SW705, for which the Göttingen miniature sire was heterozygous. Marker SW705 was at the
2249
QTL for the number of vertebrae in pigs
peak position of the plotting of F-ratios, and all alleles
of two Meishan grand dams were different from those
of the Göttingen miniature sire. Whether the sire was
homozygous or heterozygous at the QTL was determined by testing four hypotheses for QTL genotype—
H0, H1, H2, and H3—wherein we assumed the genotypes of the Göttingen miniature sire to be q/q in H0,
Q/Q in H1, Q/q in H2, and Q1/Q2 in H3. In contrast,
we assumed the QTL genotype of the Meishan dam to
be q/q in all hypotheses. In accordance with each of
these hypotheses, we inferred the QTL genotype of each
F2 individual from the genotype at SW705. We applied
linear models assuming the additive allelic effect without dominance for a QTL at position SW705 to the number of vertebrae of F2 individuals corresponding to each
of the hypotheses. Under H0, all individuals in the F2
had a genotype q/q at QTL. Under H1 and H2, three
possible genotypes—Q/Q, Q/q, and q/q—existed in the
F2, and their effects were assumed to be a, 0, and −a,
respectively. Under H3, there were six QTL genotypes—Q1/Q1, Q2/Q2, Q1/Q2, Q1/q, Q2/q, and q/q—for
which we assumed the genotypic effects to be 2a, 2b, a
+ b, a, b, and 0, respectively. We evaluated heterozygosity at the QTL for the Göttingen miniature sire on the
basis of the comparisons of goodness-of-fit in models
between H1 and H3. We measured this by log-likelihood
ratio test statistics obtained from the maximum likelihood under each of the two hypotheses to be compared,
assuming normal distributions for the number of vertebrae of F2 individuals with means corresponding to
their QTL genotypes. If H3 (Q1/Q2) showed a better (P
< 0.05) model fit than H1 (Q/Q), then we judged the
sire to be heterozygous at the QTL. The significance
level of log-likelihood ratio test statistic was determined
by a permutation test of 10,000 repetitions.
Evaluation of Heterozygosity of QTL on SSC7
in Parental European Pigs in Experimental Families
We searched for heterozygous microsatellite markers
around the QTL region in each parental European pig.
To examine the heterozygosity of the QTL on SSC7, we
used one of the heterozygous markers for each parental
European pig to infer the inheritance of each of two
homologous alleles by the F1 and F2 offspring. We determined whether each parental European pig was homozygous or heterozygous by the same statistical testing used to evaluate the heterozygosity of the QTL on
SSC1 in the Göttingen miniature sire.
Half-Sib Analysis of Sires in a Large White Population
on SSC7 for the Number of Vertebrae
We scored the number of vertebrae in 896 individuals
produced from a Large White population. The population was derived from 10 sires and 65 dams as founder
individuals and was bred in a closed population for
seven generations (1987 to 1993) with 8 to 11 sires and
28 to 36 dams in the Livestock Research Institute of
Tokushima prefecture in Japan. Using the microsatellite markers SW147 (90.1 cM in the USDA map), SW252
(99.4 cM), and S0115 (102.2 cM), which were located
on SSC7 near the QTL for vertebral number, we genotyped the 896 progeny, as well as their 25 sires and 69
dams. Haplotypes consisting of these three microsatellite markers on each homologous chromosome of the
sires were reconstructed by using genotype data of the
sires, dams, and their progeny. Among 896 progeny,
786 derived from 12 sires and 24 dams were classified
with certainty into two groups based on the haplotype
(left or right) inherited. We compared the average number of vertebrae in two groups of progeny for each sire
and used t-tests to evaluate the significance of the differences. Sires were considered to be heterozygous (Q/
q) at the QTL when differences (P < 0.05) were detected.
Moreover, to detect the sires homozygous for the QTL,
we performed Z-tests following the manner of Nezer et
al. (2003). The Q-to-q substitution effect was set at 0.43,
which was derived from the value 0.43 ± 0.10 calculated
from nine heterozygous sires. Sires were judged to be
homozygous when Z < −2.
Interaction of Two QTL on SSC1 and SSC7
for Number of Vertebrae
By using the JD family, in which the grand sire and
the five grand dams were fixed for the alternative alleles
of the QTL on both SSC1 and SSC7, a two-dimensional
search was carried out on SSC1 and SSC7 for testing
hypotheses H0 (no epistasis exists) vs. H1 (epistasis
exists). The numbers of microsatellite markers on SSC1
and SSC7 were eight and nine, respectively. A test statistic to detect epistasis was n × log (RSS0/RSS1), where
RSS0 (RSS1) is RSS (residual sum of squares) under
H0 (H1), and n is the sample size. A threshold for this
statistic was determined by a permutation test with
1,000 repetitions.
Results
Genome Scanning of QTL for the Number of Vertebrae
We performed genome scanning of QTL for the number of vertebrae in each of three swine families (MG,
ML, and LWb). In all three families, a QTL was detected
in a region around SW705 on SSC1qter (Figure 1a;
Table 2). In the ML and LWb families, another QTL
for the number of vertebrae was detected in a region
around SW252 on SSC7 (Figure 1b; Table 3). To examine the existence of the QTL in the other six families,
interval mapping analyses on SSC1 and SSC7 were performed.
QTL on SSC1 for the Number of Vertebrae
In the other six families (JW, BC, MW, LM, JD, and
WWb), QTL also were detected in a region of SSC1qter
(Table 2). In all families, the alleles from the European
2250
Mikawa et al.
pigs, at 0.44 to 0.69 (Table 2). The proportion of phenotypic variance in number of vertebrae explained by the
QTL also was less in the MG family (12.3%) than in
other families (26.0 to 42.6%).
Heterozygosity of QTL on SSC1 of the Göttingen
Miniature Sire in the MG Family
The Göttingen miniature sire had a smaller estimated effect at the QTL than those detected in other
families. It is possible that we underestimated the allelic effect of the QTL for the sire when the sire was
assumed to be homozygous at the QTL in the interval
mapping analysis. Therefore, we investigated the heterozygosity of the QTL of the Göttingen miniature sire
using the microsatellite marker SW705 (Table 4). Hypothesis testing revealed that all models of H1 (Q/Q),
H2 (Q/q), and H3 (Q1/Q2) gave better (P < 0.01) model
fits than did model H0 (q/q). When models of H1 (Q/Q)
and H3 (Q1/Q2) were compared, Model H3 (Q1/Q2) gave
a better (P < 0.01) model fit than did Model H1 (Q/Q).
Therefore, we judged the QTL to be heterozygous in
the Göttingen miniature sire. Heterozygosity of this
QTL was not detected either for the Meishan pigs in
this family or for the other European pigs used as grand
sires or grand dams in other families (data not shown).
QTL on SSC7 for the Number of Vertebrae
Figure 1. Plots of F-ratio from least squares interval
mapping analysis (Haley et al., 1994) in a) SSC1 and b)
SSC7. The x-axis indicates the relative position in the
USDA linkage map. The y-axis represents the values of
the F-ratio. 䊊 = ML (Meishan × Landrace), 䊉 = LWb
(Landrace × Japanese wild boar), and 〫 = MG (Meishan
× Göttingen miniature) families. Triangles (䉭; ▲) on the
x-axis indicate the positions of microsatellite markers.
Uppermost triangles represent the ML family. Triangles
positioned in the middle row represent the LWb family,
and lowermost triangles represent the MG family. On
SSC1, the positions of microsatellite marker SW705 are
indicated by closed triangles. On SSC7, the positions of
microsatellite markers SW147, SW252, and S0115 are indicated by closed triangles, from left to right.
breed pigs had a significant additive effect, but no significant dominance effect, increasing the number of vertebrae. In the MG family, the estimated additive effect
of the alleles from the Göttingen miniature sire was
lower (at 0.32) than those from other European breed
By genome scanning, we detected another QTL for
the number of vertebrae in a region around microsatellite marker SW252 on SSC7 in the ML (1% genomewide) and LWb (5% genome-wide) families but not in
the MG family. In the ML family, estimates of additive
effects of the QTL on SSC1 and on SSC7 in the Landrace
sire were similar at 0.59 and 0.68, respectively (Tables
2 and 3). Conversely, in the LWb family, the additive
effect of the QTL on SSC7 in the Landrace dams was
estimated to be 0.28, which was less than one-half that
on SSC1 (0.63). In both families, significant dominance
effects were not observed. In the analysis of SSC7 with
interval mapping, significant effects also were detected
around the same region on SSC7 in the JW (5% genomewide) family and in the BC, MW, and JD families (1%
genome-wide), whereas no significant effects were detected in the LM and WWb families (Table 3) or in the
MG family.
Heterozygosity of QTL on SSC7 of European
Parental Pigs in Experimental Families
As a reason for not detecting the significant effects
on SSC7 in some families, we considered that the allele
that increases the number of vertebrae was not fixed
in the European parental pigs in the experimental families. Therefore, heterozygosity of the QTL of the European parental pigs was examined in the same way that
we tested the QTL on SSC1 of the Göttingen miniature
sire in the MG family. In the LWb family, two Landrace
female pigs (L1 and L2) had been used as grand dams.
2251
QTL for the number of vertebrae in pigs
Table 2. Summary of QTL on SSC1 for the number of vertebrae
Familya
Nearest marker
to peak positionb
F-ratio
Average No.
of vertebrae
Additive
effectc
Dominance
effectc
Ratio of QTL
variance, %d
MG
ML
LWb
JW
BC
MW
LM
JD
WWb
SW705 (122.6)
SW705 (122.6)
SW705 (122.6)
SW705 (122.6)
S0112 (121.3)
S0112 (121.3)
SW705 (122.6)
SW705 (122.6)
SW705 (122.6)
12.2**
35.3**
28.8**
35.2**
29.9**
27.4**
22.8**
104.5**
101.4**
20.15
20.62
20.09
20.15
20.55
20.56
20.36
20.06
20.04
0.32
0.59
0.63
0.59
0.54
0.69
0.44
0.53
0.58
0.01
−0.07
−0.02
0.18
−0.15
0.11
0.20
0.15
0.24
12.3
34.3
31.9
42.6
34.0
41.3
26.0
28.0
38.4
a
Genome scanning was performed in the MG, ML, and LWb families. Interval mapping on SSC1 was
performed in the JW, BC, MW, LM, JD, and WWb families. See Table 1 for definition of families.
b
Positions (cM) in the USDA linkage map are presented in parentheses.
c
A positive value means that the European allele type has a positive effect.
d
The proportion of phenotypic variance explained by the QTL relative to the total variance.
**Significant at genome-wide 1% level. For the JW, BC, MW, LM, JD, and WWb families, the genomewide significance threshold was estimated from the chromosome-wide significance threshold following the
method of de Koning et al. (1999).
Analysis with microsatellite marker SW252 suggested
that one Landrace grand dam (L1) was heterozygous
at the QTL on SSC7 (Table 5). One allele had the effect
of increasing the number of vertebrae by 0.64, an
amount similar to that detected in the ML family (0.68).
For the other Landrace grand dam (L2), significant effects of the QTL were not detected in any models. In
the other families, we also analyzed the variation in
the QTL on SSC7 in the European parental pigs, using
heterozygous microsatellite markers near the QTL (Table 5). As a result, it was judged that the Large White
male pig in the JW family, the Landrace female in the
LM family, and one (W1) of the Large White females
in the WWb family were heterozygous for the QTL on
SSC7. In the ML, BC, MW, and JD families, the QTL
on SSC7 was judged to be homozygous in the European
parental pigs, and all European alleles had significant
effects in increasing the number of vertebrae. For the
Göttingen miniature male pig in the MG family and
two (W2 and W3) of the Large White females in the
WWb family, significant effects of the QTL were not
detected on SSC7 in any models.
Variation in QTL on SSC7 in a Commercial
Large White Population
In European pigs such as Landrace, Large White,
and Duroc, the number of vertebrae are not fixed and
may vary from 21 to 23. Indeed, when we scored the
number of vertebrae in a total of 896 progeny derived
from 25 sires in a Large White population, 303 pigs
(33.8%) had 21 thoracic and lumbar vertebrae, 569
(63.5%) had 22 such vertebrae, and 24 pigs (2.7%) had
23; the average total number of thoracic and lumbar
vertebrae was 21.69. To investigate whether this variation in the number of vertebrae was due to variation
in the QTL on SSC7, we performed a half-sib analysis
using the microsatellite markers SW147, SW252, and
Table 3. Summary of QTL on SSC7 for the number of vertebrae
Familya
Nearest marker
to peak positionb
F-ratio
Average No.
of vertebrae
Additive
effectc
Dominance
effectc
Ratio of QTL
variance, %d
MG
ML
LWb
JW
BC
MW
LM
JD
WWb
SW1122 (82.3)
SW252 (99.4)
SW252 (99.4)
S0115 (102.2)
SW252 (99.4)
SW147 (90.1)
SW147 (90.1)
SW252 (99.4)
SW252 (99.4)
3.0ns
40.2**
8.4*
10.8*
13.7**
15.7**
4.4ns
139.6**
7.3ns
20.15
20.62
20.09
20.15
20.55
20.56
20.36
20.06
20.04
−0.14
0.68
0.28
0.37
0.41
0.57
0.23
0.59
0.20
0.16
−0.12
0.07
−0.12
−0.12
0.15
−0.04
0.16
0.04
4.6
37.3
8.2
13.4
19.1
28.4
7.2
34.2
4.2
a
Genome scanning was performed in the MG, ML, and LWb families. Interval mapping on SSC7 was
performed in the JW, BC, MW, LM, JD, and WWb families. See Table 1 for definition of families.
b
Positions (cM) in the USDA linkage map are presented in parentheses.
c
A positive value means that the European allele type has a positive effect.
d
The proportion of phenotypic variance explained by the QTL relative to the total variance.
*Significant at the genome-wide 5% level; **significant at the genome-wide 1% level; ns = not significant.
For the JW, BC, MW, LM, JD, and WWb families, the genome-wide significance threshold was estimated
from the chromosome-wide significance threshold following the method of de Koning et al. (1999).
2252
Mikawa et al.
Table 4. Evaluation of heterozygosity of QTL on SSC1
for the number of vertebrae of the sire in MG familya
Hypothesis
testing
H0
H0
H0
H1
Additive
effects
in model
LRT
statisticb
(q/q) vs. H1 (Q/Q)
(q/q) vs. H2 (Q/q)
(q/q) vs. H3 (Q1/Q2)
(Q/Q) vs. H3 (Q1/Q2)c
24.4**
50.8**
53.3**
28.9**
0.32/0.32
0.47/0.00
0.52/0.11
a
Genotypes of SW705, which was at peak position (distance = 0
cM), were used for analysis. See Table 1 for definition of families.
b
Log-likelihood ratio test statistic was calculated by 2 × log([likelihood of Hi]/[likelihood of Hj]), where i = 0, 1; j = 1, 2, 3.
c
Calculated by ([H0 vs. H3] − [H0 vs. H1]).
**P < 0.01 based on a permutation test of 10,000 repetitions.
S0115, located on SSC7 near the QTL (Table 6). As a
result of analysis using 786 progeny derived from 12
sires, which were classified with certainty into two
groups based on the haplotype, nine sires seemed to
be heterozygous for the QTL. One sire seemed to be
homozygous, and it was not clear whether the others
were homozygous. We performed the same analysis for
the QTL on SSC1, but variation in the QTL was not
detected (data not shown).
Interaction of Two QTL on SSC1 and SSC7
for Number of Vertebrae
In the JD family, which was the largest family in
this study, the grand sire and the five grand dams were
fixed for the alternative alleles of the QTL on both SSC1
and SSC7. Using this family, we analyzed the interaction between the two QTL. A two-dimensional search
was performed on SSC1 and SSC7 to test the hypothe-
ses H0 (no epistasis present) vs. H1 (epistasis present).
No significant epistatic interaction between the two
QTL was observed (data not shown). The average number of vertebrae of F2 progeny belonging to nine categories, as classified by the genotypes of the two QTL,
showed clearly that the two QTL had an additive effect
and acted independently without interaction (Figure
2). Substitution of four Asian alleles with four European
alleles in the two QTL explained the increase in the
number of vertebrae by an average of more than two.
Discussion
In a previous study, we identified a QTL on SSC1
qter that affected the number of vertebrae in a threegeneration family derived from a cross between Meishan and Göttingen miniature pigs. In the present study,
we found that the Göttingen miniature sire was heterozygous in the QTL on SSC1 based on the comparison
between goodness of model fitting for H1 (Q/Q) and H3
(Q1/Q2). Moreover, the likelihood ratio of Model H2 (Q/
q) to Model H3 (Q1/Q2) can be calculated as 2.5 from
Table 4, considering the difference between log-likelihood ratio test statistics of H2 vs. H0 (50.8) and H3 vs.
H0 (53.3), which is not significant. Therefore, it was
inferred that the genotype of the sire is Q/q. We speculate that the allele that had this effect was derived from
a European breed and that the other was derived from
Asian breeds as part of the process of breeding Göttingen miniature pigs, which were constructed by using
European and Asian breeds (Porter, 2002). When the
QTL genotype of the sire was assumed to be heterozygous (Q/q), the proportion of phenotypic variance in
number of vertebrae explained by the QTL was calculated as 24.2%. In the experimental families in this
Table 5. Evaluation of heterozygosity of QTL on SSC7 for the number of vertebrae of the European parent pigs in
experimental families
LRTb
Family
Parent
of pig
Microsatellite
marker used
for analysisa
(q/q vs. Q/Q)
H0 vs. H1
(q/q vs. Q/q)
H0 vs. H2
(q/q vs. Q1/Q2)
H0 vs. H3
(Q/Q vs. Q1/Q2)
H1 vs. H3c
MG
ML
G
L
SW252 (99.4)
SW252 (99.4)
1.0
65.6**
4.7
39.2**
4.7
65.8**
3.7
0.2
nsd
Q/Q
—
0.68/0.68
LWb
L1
L2
SW252 (99.4)
SW252 (99.4)
5.6
5.2
13.6**
7.4
14.0**
7.5
8.4*
2.3
Q1/Q2
ns
0.64/−0.09
—
JW
BC
MW
LM
JD
W
B
W
L
D
S0115 (102.2)
SW252 (99.4)
SW147 (90.1)
SW147 (90.1)
SW252 (99.4)
19.8**
19.1**
27.9**
8.5*
231.7**
31.8**
16.3**
16.9**
20.4**
144.0**
32.3**
19.7**
28.1**
20.6**
233.0**
12.5**
0.6
0.2
12.1**
1.3
Q1/Q2
Q/Q
Q/Q
Q1/Q2
Q/Q
0.41/−0.09
0.38/0.38
0.56/0.56
0.55/0.05
0.59/0.59
WWb
W1
W2
W3
SW252 (99.4)
SW252 (99.4)
SW252 (99.4)
29.5**
0.6
0.02
40.8**
3.1
0.02
41.4**
3.1
0.02
11.9**
2.5
0.00
Q1/Q2
ns
ns
0.67/0.12
—
—
a
Model
of QTL
Positions in the USDA linkage map are presented in parentheses. See Table 1 for definition of families.
Log-likelihood ratio test statistic was calculated by 2 × log ([likelihood of Hi]/[likelihood of Hj]), where i = 0, 1; j = 1, 2, 3.
c
Calculated by ([H0 vs. H3] − [H0 vs. H1]).
d
ns = no significant effect was detected.
*P < 0.05 and **P < 0.01, respectively, based on a permutation test of 10,000 repetitions.
b
Additive
effect
in model
2253
QTL for the number of vertebrae in pigs
Table 6. Half-sib analysis of QTL on SSC7 in a Large White population
Average No. of vertebrae (No. of offspring)
Sire
P1
P2
P3
P4
P5
P6
P7
P8
P9
P10
P11
P12
Left homologue
inherited
21.43
21.55
21.49
21.42
21.46
21.18
21.50
21.70
21.68
21.91
21.55
21.68
(81)
(49)
(55)
(47)
(24)
(11)
(12)
(30)
(25)
(23)
(11)
(25)
Right homologue
inherited
21.95
22.08
21.86
21.95
21.91
21.67
21.93
22.00
21.90
21.94
21.67
21.81
(83)
(51)
(44)
(40)
(33)
(9)
(15)
(20)
(30)
(18)
(24)
(26)
QTL
q/Q
q/Q
q/Q
q/Q
q/Q
q/Q
q/Q
q/Q
q/Q
Q/Q or q/qa
ndb
nd
Significance
level
**
**
**
**
**
*
**
*
*
a
Homozygosity was judged by Z-score < −2. The Z-score was calculated by the method of Nezer et al.
(2003); the Q to q allele substitution effect was 0.43.
b
nd = could not be determined.
*P < 0.05; **P < 0.01.
study, all of the alleles of European breeds at the QTL
on SSC1 had the effect of increasing the number of
vertebrae, and Model H3 (Q1/Q2) did not give a significantly better model fit than did Model H1 (Q/Q; data not
shown). Although we analyzed only a limited number of
samples, it is possible that the allele increasing the
number of vertebrae at the QTL on SSC1 has already
spread widely in commercial pigs.
On SSC7 we also detected another QTL for the number of vertebrae; this QTL was located at a similar
Figure 2. Average number of vertebrae in F2 progeny
in each category as classified by the genotypes of the two
QTL on SSC1 and SSC7. In accordance with the genotypes
of the two QTL, 528 F2 progeny in the JD (Jinhua × Duroc)
family were classified into nine categories. Standard deviations are shown by vertical lines. The letters (a, b, c, or
d) indicate mean separation; average number of vertebrae
of F2 progeny in categories that do not have a common
letter differ (P < 0.05).
position to that reported by Sato et al. (2003). In this
study, we attempted to evaluate the heterozygosity of
the QTL on SSC7 for the Göttingen miniature and 11
European pigs used as parents of F2 experimental families and found that four European parent pigs, one each
of the LWb, JW, LM, and WWb families, were heterozygous (Q1/Q2; Table 5) in the QTL. When models of
H2 (Q/q) and H3 (Q1/Q2) were compared for the four
European parent pigs, log-likelihood ratio test statistics
of H2 vs. H3 ([H3 vs. H0] − [H2 vs. H0]) were calculated
to be 0.4, 0.5, 0.2, and 0.6 for L1 of the LWb family, W
of JW, L of LM, and W1 of WWb, respectively, and not
significant. Thus, it was inferred that genotypes of the
QTL for the four pigs were Q/q and that some of the
European alleles at this QTL had the effect of increasing the number of vertebrae. The variation of this QTL
on SSC7 also was detected in a Large White population.
Although a limited number of samples were analyzed,
it is possible that the allele that increases the number of
vertebrae was only recently introduced into European
breeds, spreading throughout the breeds along with
breeding for increased body length, and is not yet fixed.
For this reason, the QTL on SSC7 is valuable in the
breeding of commercial pigs. A possible reason for
the variation in the QTL on SSC7 in commercial pig
breeds is that an unknown locus affecting productivity
or meat quality is linked to the QTL and prevents fixation of the allele that increases the number of vertebrae.
Many QTL have been identified on SSC7. Among them,
QTL for backfat thickness and growth rate are located
30 cM away from the QTL for number of vertebrae, and
Asian (Meishan) alleles have been reported to make the
backfat thinner and growth faster (Rohrer and Keele,
1998; de Koning et al., 1999; Rohrer, 2000; Bidanel et
al., 2001). In Asian pigs used for breeding, Asian alleles
at QTL for backfat thickness and growth rate would
have been selected for. At the same time, Asian alleles
at QTL for the number of vertebrae, which would not
2254
Mikawa et al.
have had the effect of increasing the number of vertebrae, might have been reintroduced unexpectedly.
In the present study, two QTL were detected on SSC1
and SSC7 that affect the number of vertebrae associated with body length. At the QTL on SSC7, variation
was found within commercial pig breeds. This information should be useful in breeding programs, such as
those that use marker-assisted selection.
Literature Cited
Bidanel, J. P., D. Milan, N. Iannuccelli, Y. Amigues, M. Y. Boscher,
F. Bourgeois, J. C. Caritez, J. Gruand, P. Le Roy, H. Lagant, R.
Quintanilla, C. Renard, J. Gellin, L. Ollivier, and C. Chevalet.
2001. Detection of quantitative trait loci for growth and fatness
in pigs. Genet. Sel. Evol. 33:289–309.
Churchill, G. A., and R. W. Doerge. 1994. Empirical threshold values
for quantitative trait mapping. Genetics 138:963–971.
de Koning, D. J., L. G. Janss, A. P. Rattink, P. A. M. van Oers, B.
J. de Vries, M. A. Groenen, J. J. van der Poel, P. N. de Groot,
E. W. Brascamp, and J. A. van Arendonk. 1999. Detection of
quantitative trait loci for back fat thickness and intramuscular
fat content in pigs (Sus scrofa). Genetics 152:1679–1690.
Haley, C. S., S. A. Knott, and J. M. Elsen. 1994. Mapping quantitative
trait loci in crosses between outbred lines using least squares.
Genetics 136:1195–1207.
Kamimura, R., S. Suzuki, S. Nozaki, H. Sakamoto, H. Maruno, and
H. Kawaida. 1996. Branching patterns in coronary artery and
ischemic areas induced by coronary arterial occlusion in the
Clawn miniature pig. Exp. Anim. 45:149–153.
King, J. W. B., and R. C. Roberts. 1960. Carcass length in the bacon
pig; Its association with vertebrae numbers and prediction from
radiographs of the young pig. Anim. Prod. 2:59–65.
Mikawa, S., T. Akita, N. Hisamatsu, Y. Inage, Y. Ito, E. Kobayashi,
H. Kusumoto, T. Matsumoto, H. Mikami, M. Minezawa, M. Miyake, S. Shimanuki, C. Sugiyama, Y. Uchida, Y. Wada, S. Yanai,
and H. Yasue. 1999. A linkage map of 243 DNA markers in
an intercross of Göttingen miniature and Meishan pigs. Anim.
Genet. 30:407–417.
Nakanishi, Y. 1981. Body measurements and some characteristics
of inbred CLAWN miniature pigs. Jpn. J. Swine Sci. 28:211–218.
Nezer, C., C. Collette, L. Moreau, B. Brouwers, J. J. Kim, E. Giuffra,
N. Buys, L. Andersson, and M. Georges. 2003. Haplotype sharing
refines the location of an imprinted quantitative trait locus with
major effect on muscle mass to a 250-kb chromosome segment
containing the porcine IGF2 gene. Genetics 165:277–285.
Nii, M., T. Hayashi, S. Mikawa, F. Tani, A. Niki, N. Mori, Y. Uchida,
N. Kanaya, M. Komatsu, and T. Awata. 2005. Quantitative trait
loci mapping for meat quality traits and muscle fiber property
in a Japanese wild boar × Large White intercross. J. Anim. Sci.
83:308–315.
Porter, V. 2002. Mason’s World Dictionary of Livestock Breeds, Types
and Varieties. 5th ed. CAB International, Wallingford, UK.
Rohrer, G. A. 2000. Identification of quantitative trait loci affecting
birth characters and accumulation of backfat and weight in a
Meishan-White Composite resource population. J. Anim. Sci.
78:2547–2553.
Rohrer, G. A., L. J. Alexander, Z. Hu, T. P. L. Smith, J. W. Keele,
and C. W. Beattie. 1996. A comprehensive map of the porcine
genome. Genome Res. 6:371–391.
Rohrer, G. A., and J. W. Keele. 1998. Identification of quantitative
trait loci affecting carcass composition in swine: I. Fat deposition
traits. J. Anim. Sci. 76:2247–2254.
Sato, S., Y. Oyamada, K. Atsuji, T. Nade, S. Sato, E. Kobayashi, T.
Mitsuhashi, K. Nirasawa, A. Komatsuda, Y. Saito, S. Terai,
T. Hayashi, and Y. Sugimoto. 2003. Quantitative trait loci analysis for growth and carcass traits in a Meishan × Duroc F2 resource population. J. Anim. Sci. 81:2938–2949.
Wada, Y., T. Akita, T. Awata, T. Furukawa, N. Sugai, Y. Inage,
K. Ishii, Y. Ito, E. Kobayashi, H. Kusumoto, T. Matsumoto, S.
Mikawa, M. Miyake, A. Murase, S. Shimanuki, T. Sugiyama, Y.
Uchida, S. Yanai, and H. Yasue. 2000. Quantitative trait loci
(QTL) analysis in a Meishan × Göttingen cross population. Anim.
Genet. 31:376–384.