Genetic association of productive and

reproductive traits with stayability in Nellore

cattle: analysis using Bayesian models
E.C.A. Rizzo1, F.R.A. Neto2, I.D.P.S. Diaz2, M.M. Dias2, R.B. Costa2,
H.T. Ventura3,4, H.N. Oliveira2 and A.J.S. Falco1

1
Escola de Medicina Veterinria e Zootecnia, Universidade Federal do Tocantins,
Araguana, TO, Brasil
Departamento de Zootecnia, Universidade Estadual de So Paulo Jlio de
Mesquita Filho, Jaboticabal, SP, Brasil
Associao Brasileira dos Criadores de Zebu, Uberaba, MG, Brasil
4
Department of Biochemistry, College of Science, King Saud University, Riyadh,
Saudi Arabia

Corresponding author: E.C.A. Rizzo

E-mail: [email protected] / [email protected]

Genet. Mol. Res. 14 (4): 14956-14966 (2015)

Received April 17, 2015
Accepted August 11, 2015
Published November 24, 2015
DOI http://dx.doi.org/10.4238/2015.November.24.3

ABSTRACT. We evaluated the genetic association of growth traits [weight

adjusted to 205 days of age (W205), 365 days of age (W365), and 550
days of age (W550); weight gain between 205 days of age and 365 days of
age (WG1) and between 365 days of age and 550 days of age (WG2)] and
reproductive traits [age at first calving (AFC); first calving interval (FCI)]
with stayability in the herd (STAY), using Bayesian inference in linear and
threshold models. We defined STAY as the probability of a cow calving
three or more times before the age of 76 months, given that she had
calved at least once. We assigned binary codes (0, failure; 1, success)
to each female. We used a sire model for analysis and formed different
contemporary groups for the investigated traits. We analyzed the results
by applying a two-trait sire model that included STAY (threshold trait) and
linear traits (W205, W365, W550, WG1, WG2, AFC, and FCI). We used

Genetics and Molecular Research 14 (4): 14956-14966 (2015) FUNPEC-RP www.funpecrp.com.br

 Genetic association of growth traits with stayability 14957

Gibbs sampling to estimate variance components and heritabilities. In all

the analyses, we found that the mean heritability estimates for STAY were of
moderate magnitude (0.20-0.25). The mean heritabilities for W205, W365,
W550, WG1, WG2, AFC, and FCI were 0.20, 0.23, 0.39, 0.08, 0.14, 0.12, and
0.11, respectively. We observed wide variation in the posterior distributions
of genetic correlations; however, with the exception of those obtained for the
reproductive traits, the mean estimates were of low magnitude. Selection for
WG2 can results in favorable correlated response in STAY.

Key words: Heritability; Longevity; Selection of females; Threshold model

INTRODUCTION

An increase in cow longevity reduces annual production costs derived from heifer
replacement and involuntary culling. Consequently, higher selection intensities can be applied to
female selection criteria in breeding programs.
Production during the life of a cow is of direct economic importance but is difficult to define.
Moreover, selection for this trait is associated with a series of problems. First, the cow needs to leave
the herd so that the trait is expressed. Therefore, the main problem related to the analysis of measures
of longevity-associated traits such as the length of productive life of a cow in the herd is the presence
of censoring, i.e., the recording of partial observations. Several animals of the contemporary groups
that are being analyzed will remain in the herd throughout the study. Hence, the data obtained from
these animals cannot be used in the analysis, and this leads to bias in the results.
Several mathematical expressions have been used to define traits that could be used to
express the total production of a cow. These traits include stayability in the herd (STAY), number
of calves produced, and calf weight (Lbo et al., 1994, 2000; Eler et al., 2006). One of the traits
most commonly used to represent the useful life of females is STAY, which is evaluated in dairy
and beef cattle herds.
STAY is defined by some authors as the probability that a cow will survive to a specific age,
given that it has the opportunity to reach this age (Hudson and Van Vleck, 1981; Marcondes et al.,
2005). Other researchers define STAY as the fraction of progeny of a certain bull that is still alive at
a particular age, given that these progeny are a part of a group that has had the opportunity to reach
this age. According to Van Vleck (1980), the advantages of STAY are that this trait is easily measured
and it represents the selection criteria of the breeder who eliminates cows based on low production.
Another advantage is that STAY is expressed at a fixed age; hence, it is not necessary to wait for
the animals to be culled from the herd before measuring the trait. However, because stayability is
measured later in life than age at first calving (AFC) and first calving interval (FCI), its use as selection
criteria could increase generation interval and reduce genetic progress on AFC and FCI.
The use of STAY in genetic evaluations has a series of limitations owing to the binary
nature of the observations. According to Smith and Allaire (1986), linear models-, which do not
accommodate censored data-are also inadequate for the analysis of binomial data because of the
violation of the assumption of normality. In contrast, threshold models for binary data have been
used efficiently for decade (Rekaya et al., 2001; Silva et al., 2003).
According to Silva et al. (2003), STAY in beef cattle represents an indicator trait of
reproductive capacity. The reason is that in beef cattle farming, cows are culled because of
reproductive failure and not based on low productive capacity. STAY is currently used an indicator

Genetics and Molecular Research 14 (4): 14956-14966 (2015) FUNPEC-RP www.funpecrp.com.br

 E.C.A. Rizzo et al. 14958

trait of reproductive capacity in various beef cattle breeding programs in Brazil. However, to the
best of our knowledge, no previous studies have examined the association of this trait with other
productive and reproductive traits, in order to indicate the effect of selection for STAY on other traits
of economic importance or to identify an indicator trait of longevity.
In the present study, we evaluated the genetic association of growth and reproductive
traits with STAY, using Bayesian inference in linear and threshold models.

MATERIAL AND METHODS

Traits and data consistency

The data used in this study refer to Nellore herds reared in the northern region of Brazil
and were provided by the Brazilian Association of Zebu Breeders which is located in Uberaba - MG,
Brazil [Associao Brasileira dos Criadores de Zebu (ABCZ)]. We studied the following traits: weight
adjusted to 205 days of age (W205), 365 days of age (W365), and 550 days of age (W550); weight
gain between standard ages [205-365 days (WG1) and 365-550 days (WG2)]; age at first calving
(AFC) first calving interval (FCI); and STAY. We defined STAY as a binary variable, attributing a value
of 1 to cows with at least three calvings before the age of 76 months and a value of 0 to cows with
fewer than three calvings before the age of 76 months. According to Formigoni et al. (2002), three
calvings cover the costs of breeding and rebreeding of a cow. We established a period of 76 months
to enable the cow to reach three calvings, because the first calving generally occurs at 32 months of
age. Only data obtained from females with STAY records were included in our analyses.
We analyzed the structure and consistency of the dataset using the Statistical Analysis
System program (SAS, 2004). The environmental factors used for the formation of the contemporary
groups varied according to the trait studied. For weights and weight gains, the contemporary groups
were formed by animals born on the same farm, in the same year and season of birth, and reared
under the same feeding regimen. For AFC and STAY, only animals born on the same farm and in
the same year and season of birth were considered. For FCI, animals were grouped according to
cows farm, year of birth and season of first calving. Contemporary groups with at least 10 animals
and offspring of a single sire were maintained. In the case of STAY, contemporary groups showing
no variation were excluded. For the remaining traits, animals with performance records having 3.5
standard deviations above or below the mean of their contemporary group were excluded. The
descriptive statistics of the dataset used are shown in Table 1.

Table 1. Descriptive statistics of the dataset used for the combined analysis of STAY (T1) and linear traits (T2).

T2 N (T1) N (T2) Mean (T2) SD (T2) P (0)

W205 8407 19,441 172.07 kg 27.48 kg 42.9%

W365 7813 15,402 227.82 kg 45.36 kg 42.7%
W550 7452 9834 278.88 kg 47.16 kg 43.5%
WG1 7813 15,514 55.33 kg 32.92 kg 42.7%
WG2 7452 9962 60.65 kg 22.61 kg 43.5%
AFC 11,055 49,174 1147 days 138 days 41.6%
FCI 11,027 27,468 494 days 113 days 41.6%

N, number of observations; SD, standard deviation; P (0), proportion of animals that did not achieve three calvings until
6 years of age; W205, weight adjusted to 205 days of age; W365, weight adjusted to 365 days of age; W550, weight
adjusted to 550 days of age; WG1, weight gain between 205 days and 365 days of age; WG2, weight gain between
365 days and 550 days of age; AFC, age at first calving; FCI, first calving interval.

Genetics and Molecular Research 14 (4): 14956-14966 (2015) FUNPEC-RP www.funpecrp.com.br

 Genetic association of growth traits with stayability 14959

Genetic-quantitative analysis

In the present study, we conducted two-trait analysis in a sire model that included STAY
(threshold trait), growth traits (W205, W365, W550, WG1, and WG2), and reproductive traits (AFC
and FCI). The matrix representation of the analysis model is:

(Equation 1)

where y, b, s, and are the vectors of observations, effects (contemporary groups), random sire
effect, and random residual effect, respectively; and X and Z, in this order, are the incidence
matrices relating the effects to the observations.
We estimated variance and covariance components using Bayesian methods based
on the marginal posterior distributions of the parameters. We considered the following a priori
distributions: uniform for the effects of contemporary groups, Gaussian for sire and residual effects,
and inverse Wishart distributions, which are inverse multivariate gamma distributions, for sire (G)
and residual (R) (co)variance components.
The threshold model used for the analysis of STAY was based on the theory initially
proposed by Sorensen (1996) and Korsgaard et al. (2003), using Bayesian inference through
Gibbs sampling in single- and multi-trait models. In these models, it is assumed that the underlying
scale shows a normal distribution, which can be described as:

(Equation 2)

where U is the underlying variable vector of order r; is the vector of the location
parameters of order s with p fixed effects (b) and k random additive genetic values (a); W is a known
2
incidence matrix of order r x s; I is an identity matrix of order r; and e is the residual variance. The
2
variable in the underlying distribution is not observable, and therefore parameterization e = 1 is
generally adopted to obtain reliability in the likelihood function (Gianola and Sorensen, 2002).
Categorical traits are determined by non-observable continuous variables in the underlying
scale and initial threshold values are established, where t1 < t2 < ... < tj 1, with t0 = and tj = +,
with j being the number of categories. The conditional probability that the response variable, yi, falls
in category j (j = 0;1) can be written as:

(Equation 3)

where yi is the response variable for the ith observation; t is the value of the threshold that defines
the categories of the response; Ui is the value of the underlying variable; ( ) is the cumulative
distribution function of a standardized normal variate; is a row incidence vector that associate
to the ith observation.
The analyses were carried out using the THRGIBBSF90 program (Misztal, 2007), in which
methods of Bayesian inference through the Gibbs sampler are implemented. The analysis was
performed using a single chain of 400,000 samples, with a conservative burn-in period of 40,000
cycles and retention of a sample for every 10 iterations. Next, convergence of the Markov chains was

Genetics and Molecular Research 14 (4): 14956-14966 (2015) FUNPEC-RP www.funpecrp.com.br

(Equation 3)
 E.C.A. Rizzo et al. 14960

confirmed using graphic (visual) inspection and the Geweke test (1992). Thus, descriptive statistics
of the samples were obtained considering a stationary chain. The genetic parameters (heritabilities
and genetic correlations) and breeding values were calculated vector by vector, and distributions
were obtained for these parameters. In view of the possibility of obtaining asymmetrical posterior
distributions of the parameters, the 95% highest posterior density intervals of the samples were
calculated. Convergence diagnostics were performed using the Bayesian Output Analysis program
(BOA) (Smith, 2005), available in the R Core Team programming package (R Core Team, 2012).

RESULTS

The convergence diagnostics revealed that the burn-in period adopted was sufficient to
obtain stationary chains; with the exception of one parameter, the z-scores were within the range
of -1.96 to +1.96 in the Geweke test (Tables 2-4). Posterior densities and trace plots for adjusted
weights, weight gains, and reproductive traits, are shown in Figures 1, 2, and 3, respectively.
The null hypothesis for residual variance of WG1 was rejected in the Geweke test (Table 3);
however, visual inspection of the sample sequence trace (Figure 2) showed that the chain could be
considered stable even in this case.

Figure 1. Posterior density (left) and trace plots (right) of heritabilities for weights adjusted for 205 days (hW205), 365 days
(hW365), and 550 days (hW550), stayability (hSTAY), genetic correlation (ra), and residual correlation (re) for each weight.

Figure 2. Posterior density (left) and trace plots (right) of heritabilities for weight gain between 205 and 365 days
(hWG1) and between 365 and 550 days (hWG2), stayability (hSTAY), genetic correlation (ra), and residual correlation (re)
for each trait.

Genetics and Molecular Research 14 (4): 14956-14966 (2015) FUNPEC-RP www.funpecrp.com.br

 Genetic association of growth traits with stayability 14961

Figure 3. Posterior density (left) and trace plots (right) of heritabilities for age at first calving (hAFC) and first calving
interval (hFCI), stayability (hSTAY), genetic correlation (ra), and residual correlation (re) for each trait.

All the posterior distributions of the parameters tended to approach a normal distribution;
however, the correlations showed wide highest posterior density intervals.
In all the analyses, the heritability estimates (Tables 2-4) for STAY were of moderate
magnitude (0.20-0.25).

Table 2. Posterior means, 95% highest posterior density intervals (95HPD), and Geweke z-scores of the
parameters estimated in two-trait analysis between stayability (STAY) and weights adjusted for 205 days (W205),
365 days (W365), and 550 days (W550) of age.

Weight Parameter* Mean 95HPD Geweke

Lower limit Upper limit

W205 s12 0.07 0.03 0.11 0.19

s22 22.32 14.86 30.35 0.69
e1 2
1.02 0.99 1.05 0.27
e22 413.82 404.90 421.90 1.12
h1
2
0.25 0.13 0.38 0.18
h 22 0.20 0.14 0.27 0.67
s 0.09 -0.23 0.39 1.79
r 0.09 0.004 0.19 -0.45
W365 s12 0.06 0.03 0.09 0.37
s2 2
50.84 33.80 69.76 -1.01
e12 1.02 0.99 1.05 -1.49
e2 2
831.38 811.70 849.90 1.01
h1
2
0.23 0.12 0.35 0.41
h 22 0.23 0.16 0.31 -1.03
s 0.07 -0.26 0.37 -0.38
r -0.01 -0.12 0.09 -1.85
W550 s12 0.06 0.03 0.10 0.36
s2 2
95.41 62.15 132.10 -0.19
e1 2
1.03 0.99 1.06 0.56
e22 886.30 859.80 911.30 -0.38
h1
2
0.23 0.12 0.36 0.35
h 22 0.39 0.26 0.52 -0.17
s 0.06 -0.27 0.38 -1.72
r 0.19 0.06 0.32 1.43

*s12, sire variance for STAY; s22, sire variance for the weight analyzed; e12, residual variance for STAY; e22, residual
variance for the weight analyzed; h12, heritability for STAY; h22, heritability for the weight analyzed; s, genetic correlation;
r, residual correlation.

Genetics and Molecular Research 14 (4): 14956-14966 (2015) FUNPEC-RP www.funpecrp.com.br

 E.C.A. Rizzo et al. 14962

Table 3. Posterior means, 95% highest posterior density intervals (95HPD), and Geweke z-scores of the
parameters estimated in two-trait analysis between stayability (STAY) and weight gain between the adjusted
ages of 205 and 365 days (WG1) and 365 and 550 days (WG2).

Weight gain Parameter Mean 95HPD Geweke

Lower limit Upper limit

WG1 s12 0.06 0.039 0.09 1.39

s22 9.23 4.74 14.15 -0.74
e12 1.02 0.99 1.05 -0.34
e22 456.17 445.50 466.40 2.29
h 12 0.23 0.11 0.35 1.41
h 22 0.08 0.04 0.12 -0.77
s 0.02 -0.38 0.41 -0.14
r 0.04 -0.08 0.16 1.22
WG2 s12 0.06 0.03 0.09 -0.54
s22 7.49 4.09 11.32 0.96
e12 1.03 0.99 1.06 0.08
e22 199.08 193.30 204.70 1.85
h 12 0.23 0.11 0.35 -0.53
h 22 0.14 0.08 0.22 0.90
s 0.19 -0.19 0.55 0.66
r 0.27 0.16 0.39 0.14

*s12, sire variance for STAY; s22, sire variance for the weight gain analyzed; e12, residual variance for STAY; e22,
residual variance for the weight gain analyzed; h12, heritability for STAY; h22, heritability for the weight gain analyzed;
s, genetic correlation; r, residual correlation.

The posterior means of the variance components and genetic parameters are shown in Table
2 (adjusted weights), Table 3 (weight gains), and Table 4 (AFC and FCI). The mean heritabilities for
W205, W365, and W550 were 0.20, 0.23, and 0.39, respectively. The mean heritabilities for weight
gains were of low magnitude (WG1, 0.08; WG2, 0.14) (Table 3). Moreover, the mean estimates of
0.12 and 0.11 obtained for AFC and FCI, respectively (Table 4) were considered low.

Table 4. Posterior means, 95% highest posterior density intervals (95HPD), and Geweke z-scores of the
parameters estimated in two-trait analysis between stayability (STAY) and the two investigated reproductive traits,
namely, age at first calving (AFC) and first calving interval (FCI).

Reproductive trait Parameter Mean 95HPD Geweke

Lower limit Upper limit

AFC s12 0.05 0.03 0.08 -0.89

s22 427.19 314.30 553.30 -0.79
e1 2
1.01 0.99 1.03 1.26
e2 2
13,361 13,180 13,520 1.81
h1
2
0.20 0.12 0.29 -0.89
h2
2
0.12 0.09 0.16 -0.85
s -0.69 -0.85 -0.51 0.47
r -0.64 -0.65 -0.62 0.29
FCI s1 2
0.06 0.034 0.090 -1.14
s2 2
251.88 165.90 342.90 -0.05
e12 1.02 0.99 1.04 -0.73
e2 8628 8476 8777
2
-0.65
h 12 0.23 0.13 0.33 -1.13
h2
2
0.11 0.07 0.15 -0.04
s -0.60 -0.83 -0.36 0.19
r -0.51 -0.54 -0.49 1.03

*s12, sire variance for STAY; s22, sire variance for the reproductive trait analyzed; e12, residual variance for STAY; e22,
residual variance for the reproductive trait analyzed; h12, heritability for STAY; h22, heritability for the reproductive trait
analyzed; s, genetic correlation; r, residual correlation.

Genetics and Molecular Research 14 (4): 14956-14966 (2015) FUNPEC-RP www.funpecrp.com.br

 Genetic association of growth traits with stayability 14963

DISCUSSION

In the present study, all the posterior distributions of the parameters (Figures 1-3) tended
to approach a normal distribution. However, the correlations showed wide highest posterior density
intervals; hence, conclusions about these parameters must be drawn with caution. According to
Lynch (1999), in comparison with univariate parameters, genetic correlations are particularly difficult
to estimate because they require accurate estimates of the parameters of which they are functions.
In all the analyses, the heritability estimates (Tables 2-4) for STAY were of moderate magnitude
(0.20-0.25). These results suggest that a considerable part of genetic variation can be attributed to the
average effects of genes, and this enables genetic gains to be obtained through selection. Similar
heritability estimates have been reported in previous studies evaluating data obtained from Nellore
cattle. Silva et al. (2006) determined an estimate of 0.22. Melis et al. (2007) obtained heritabilities
of 0.25, 0.22, and 0.28 for STAY at 5 years of age, 6 years of age, and 7 years of age, respectively.
However, Buzanskas et al. (2010) reported a lower heritability (0.03) for this trait. This wide range in
heritability estimates can be attributed to the different models used (sire or animal model) or the different
definitions of STAY adopted, as well as to the different genetic constitutions of the populations studied.
The heritabilities for W205, W365, and W550 were 0.20, 0.23, and 0.39, respectively.
An increasing trend in heritability was previously observed in longitudinal studies of weights
(Albuquerque and Meyer, 2001; Boligon et al., 2010). The heritabilities obtained in our present
study are similar to those reported by Giannotti et al. (2005) for weights from birth to 550 days of age
(0.23-0.31). However, they are lower than those reported by Siqueira et al. (2003), who analyzed
weights at different ages (120, 455, and 550 days) and estimated heritabilities ranging from 0.29
to 0.58 for these traits. Faria et al. (2007) obtained higher heritability estimates for weights at 365
days (0.49) and 450 days (0.52) of age.
In the present study, the mean heritabilities for weight gains were of low magnitude (WG1,
0.08; WG2, 0.14) (Table 3). These estimates are lower than those obtained by Paneto et al.
(2002) for Nellore cattle, using the REML method in single-trait analysis. These authors estimated
heritabilities for WG120-240, WG240-365, WG365-455, and WG455-550 of 0.32, 0.16, 0.21, and
0.23, respectively. However, the heritability for WG2 obtained in the present study is higher than
those reported by Cucco (2008) for Brown Swiss cattle-0.12 (WG120-205), 0.12 (WG205-365),
0.13 (WG365-450), and 0.05 (WG450-365).
The heritability for AFC (0.12) obtained in the present study is similar to those reported by
Dias et al. (2004), Silva et al. (2005), and Boligon et al. (2007, 2008) for Nellore cattle. Reproductive
traits measured in Nellore females were previously reported to show low heritabilities, because
these traits are strongly influenced by environmental factors and non-additive genetic effects.
However, heritability estimates for AFC of moderate to high magnitude-ranging from 0.28 to 0.40-
have been demonstrated (Mercadante et al., 2000; Bertazzo et al., 2004; Faria et al., 2008).
The heritability for FCI obtained in the present study was 0.11 and was considered low. On
the other hand, Campello et al. (1999) and Silveira et al. (2004) studied calving intervals considering
various measures per animal and estimated moderate and high coefficients of heritability (0.32 and
0.42, respectively). According to these authors, low coefficients of heritability are obtained when
only the FCI or FCIs are evaluated; the main reason is that young cows are more vulnerable to
nutritional stress, and this impairs the identification of genetic differences. The traits AFC and FCI
are generally of fundamental importance for the productivity of a herd. Gains in these traits obtained
through genetic improvement will be permanent, and this justifies investment in the selection of
genetically superior animals.

Genetics and Molecular Research 14 (4): 14956-14966 (2015) FUNPEC-RP www.funpecrp.com.br

 E.C.A. Rizzo et al. 14964

We observed wide variation in the posterior distributions of genetic correlations between

STAY and the other investigated traits (Figures 1-3); however, with the exception of those obtained
for the reproductive traits, the mean estimates were of low magnitude. Values indicating a low
association between weight and longevity traits have been reported previously. Buzanska et al.
(2010) determined a genetic correlation of -0.09 between STAY and weight at 420 days of age in
Canchim cattle. Silva et al. (2006) reported a positive genetic correlation (0.15) between STAY and
weight at 550 days of age in Nellore cattle.
Based on the definition of STAY, the negative association of AFC and FCI with STAY is
not surprising; in other words, the earlier an animal initiates breeding and the shorter its calving
intervals, the higher will be the probability of success (at least three calvings before the age of 6
years). Silva (2012) obtained values of the same order of magnitude for Gyr cattle, by applying
two different definitions of STAY at 48 months (-0.55) and 60 months of age (-0.54). However,
contrary to the present study, the author reported positive correlations between STAY and FCI; this
discrepancy was probably caused by the different definitions of STAY used.
Considering the magnitude of heritability for STAY obtained in the present study, genetic gain by
direct selection would be expected. However, this trait is measured at an advanced age, and therefore
the annual genetic gain may be compromised because of an increase in the generation interval. Hence,
an alternative means of genetic gain would be indirect selection through traits that are measured at
younger ages, and which are correlated with STAY. Given the genetic correlations, the most appropriate
option for obtaining correlated gain would be to select animals for AFC-even if the accuracy of selection
is lower-based on the magnitude of heritability. However, different definitions of criteria for the cows
initiation of breeding exist on different farms, and therefore the use of AFC as a selection criterion must
be treated with caution. Furthermore, AFC is obtained earlier than STAY but is measured only after the
animals have initiated breeding. Moreover, this trait is only expressed in females.
Most genetic breeding programs apply weight traits as selection criteria, and therefore the
use of W550 and WG2 would be a more viable option. First, both traits are observed at younger ages
than AFC and therefore they may contribute to a decrease in the generation interval. Considering
the factors that influence correlated genetic gain, selection for WG2 would be more interesting
based on the higher genetic correlation with STAY. However, W550 presents higher heritability, and
therefore this trait would permit higher gains owing to the greater accuracy of selection.

Conflicts of interest

The authors declare no conflict of interest.

ACKNOWLEDGMENTS

The authors thank the Associao Brasileira dos Criadores de Zebu (ABCZ) for providing
the dataset and Coordenao e Aperfeioamento de Pessoal de Nvel Superior (CAPES) for
financial support.

REFERENCES
Albuquerque LG and Meyer K (2001). Estimates of covariance functions for growth from birth to 630 days of age in Nelore
cattle. J. Anim. Sci. 79: 2776-2789.
Bertazzo RP, Freitas RTF, Gonalves TM, Pereira IG, et al. (2004). Parmetros genticos de longevidade e produtividade de
fmeas da raa Nelore. Rev. Bras. Zootec. 33: 1118-1127.

Genetics and Molecular Research 14 (4): 14956-14966 (2015) FUNPEC-RP www.funpecrp.com.br

 Genetic association of growth traits with stayability 14965

Boligon AA, Albuquerque LG, Mercadante MEZ and Lbo RB (2010). Study of relations among age at first calving, average
weight gains and weights from weaning to maturity in Nellore cattle. Rev. Bras. Zootec. 39: 746-751.
Boligon AA, Albuquerque LG and Rorato PRN (2008). Associaes genticas entre pesos e caractersticas reprodutivas em
rebanhos da raa Nelore. Rev. Bras. Zootec. 37: 596-601.
Boligon AA, Rorato PRN and Albuquerque LG (2007). Correlaes genticas entre medidas de permetro escrotal e
caractersticas produtivas e reprodutivas de fmeas da raa Nelore. Rev. Bras. Zootec. 36: 565-571.
Buzanskas ME, Grossi DA, Baldi F, Barrozo D, et al. (2010). Genetic associations between stayability and reproductive and
growth traits in Canchim beef cattle. Livest. Sci. 132: 107-112.
Campello CC, Martins Filho R and Lbo RNB (1999). Intervalo de partos e fertilidade real em vacas nelore no estado do
Maranho. Rev. Bras. Zootec. 28: 474-479.
Cucco DC (2008). Estimativa de parmetros genticos para caractersticas de crescimento e permetro escrotal na raa Pardo
Suo Corte. Masters thesis, Faculdade de Zootecnia e Engenharia de Alimentos, Universidade Estadual de So Paulo,
Pirassununga.
Dias LT, El Faro L and Albuquerque LG (2004). Estimativas de herdabilidade para idade ao primeiro parto de novilhas da raa
Nelore. Rev. Bras. Zootec. 33: 97-102.
Eler JP, Ferraz JBS, Balieiro JCC, Mattos EC, et al. (2006). Genetic correlation between heifer pregnancy and scrotal
circumference measured at 15 and 18 months of age in Nellore cattle. Genet. Mol. Res. 5: 569-580.
Faria CU, Magnabosco CU, Albuquerque LG, Los Reyes A, et al. (2008). Abordagem bayesiana e freqentista em anlise
gentica tricarcter para crescimento e reproduo de bovinos Nelore. Cinc. Anim. Bras. 9: 598-607.
Faria CU, Magnabosco CU, Reyes A, Lbo RB, et al. (2007). Bayesian inference in a quantitative genetic study of growth traits
in Nelore cattle (Bos indicus). Genet. Mol. Biol. 30: 545-551.
Formigoni IB, Silva JAIIV, Brumatti RC, Ferraz JBS, et al. (2002). Economic aspects of stayability as selection criterion in beef
cattle industry in Brazil. Proceedings of the 7th World Congress on Genetics Applied to Livestock Production, Montpellier,
CD-ROM. Section 2, Communication 02-62.
Geweke J (1992). Evaluating the accuracy of sampling-based approaches to the calculation of posterior moments. In: Bayesian
Statistics. 4th edn. (Bernardo JM, Berger JO, Dawid AP, and Smith AFM, eds.). Oxford University Press, New York,
169-193.
Giannotti JDG, Packer IU and Mercadante MEZ (2005). Meta-anlise para estimativas de herdabilidade para caractersticas
de crescimento em bovinos de corte. Rev. Bras. Zootec. 34: 1173-1180.
Gianola D and Sorensen D (2002). Likelihood, Bayesian, and MCMC methods in quantitative genetics. 2nd edn. Statistics for
Biology and Health. Springer-Verlag, New York.
Hudson GFS and Van Vleck LD (1981). Relationship between production and stayability in Holstein cattle. J. Dairy. Sci. 64:
2246-2250.
Korsgaard IR, Lund MS, Sorensen D, Gianola D, et al. (2003). Multivariate Bayesian analysis of Gaussian, right censored
Gaussian, ordered categorical and binary traits using Gibbs sampling. Genet. Sel. Evol. 35: 159-183.
Lbo RNB, Madalena FE and Vieira AR (2000). Means of published genetic parameter estimates for cattle raised on tropical
climate. Anim. Breed. Abstr. 68: 433-462.
Lbo RB, Reyes A, Ferraz JBS, Bezerra MEZ, et al. (1994). Bivariate animal model analysis of growth weights and scrotal
circumference of Nellore cattle in Brazil. In: World Congress on Genetics Applied to Livestock Production, 5, Guelph, 199-202.
Lynch M (1999). Estimating genetic correlations in natural populations. Genet. Res. 74: 255-264.
Marcondes CR, Paneto JCC, Bezerra LAF and Lbo RB (2005). Estudo de definio alternativa da probabilidade de
permanncia no rebanho para a raa Nelore. Rev. Bras. Zootec. 34: 1563-1567.
Melis MH Van, Eler JP, Oliveira HN, Rosa GJM, et al. (2007). Study of stayability in Nellore cows using a threshold model. J.
Anim. Sci. 85: 1780-1786.
Mercadante MEZ, Lbo RB and Oliveira HN (2000). Estimativas de (co)varincias entre caractersticas de reproduo e de
crescimento em fmeas de um rebanho Nelore. Rev. Bras. Zootec. 29: 997-1004.
Misztal I (2007). BLUPF90 family of programs. Available at http://nce.ads.uga.edu/~ignancy/ newprograms.html.
Paneto JCC, Lemos DC, Bezerra LAF, Martins Filho R, et al. (2002). Estudo de caractersticas quantitativas de crescimento
dos 120 aos 550 dias de idade em gado Nelore. Rev. Bras. Zootec. 31: 668-674.
R Core Team (2012). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna,
Austria. Available at http://www.R-project.org/.
Rekaya R, Weigel KA and Gianola D. (2001). Threshold model for misclassified binary responses with applications to animal
breeding. Biometrics 57: 1123-1129.
SAS Institute Inc. (2004). SAS OnlineDoc 9.1.3. SAS Institute Inc., Cary.
Silva JAIIV, Dias LT and Albuquerque LG (2005). Estudo gentico da precocidade sexual de novilhas em um rebanho Nelore.
Rev. Bras. Zootec. 34: 1568-1572.

Genetics and Molecular Research 14 (4): 14956-14966 (2015) FUNPEC-RP www.funpecrp.com.br

 E.C.A. Rizzo et al. 14966

Silva JAIIV, Eler JP, Ferraz JBS, Golden BL, et al. (2003). Heritability estimate for stayability in Nelore cows. Livest. Prod. Sci.
79: 97-101.
Silva JAIIV, Formigoni IB, Eler JP and Ferraz JBS (2006). Genetic relationship among stayability, scrotal circumference and
post-weaning weight in Nelore cattle. Livest. Sci. 99: 51-59.
Silva RMO (2012). Estimativas de parmetros genticos para habilidade de permanncia no rebanho e suas associaes com
caractersticas de interesse econmico em vacas da raa Gir Leiteiro. Masters thesis, Universidade Federal de Alagoas,
Rio Largo.
Silveira JC, McManus C, Mascioli AS, Silva LOC, et al. (2004). Fatores ambientais e parmetros genticos para caractersticas
produtivas e reprodutivas em um rebanho nelore no estado do Mato Grosso do Sul. Rev. Bras. Zootec. 33: 1432-1444.
Siqueira RLPG, Oliveira JA, Lbo RB, Bezerra LAF, et al. (2003). Anlise da variabilidade gentica aditiva de caractersticas
de crescimento na raa Nelore. Rev. Bras. Zootec. 32: 99-105.
Smith BJ (2005). Bayesian Output Analysis program (BOA). Version 1.1 Users Manual. Available at [http://www.public-health.
uiowa.edu/boa/BOA.pdf].
Smith SP and Allaire FR (1986). Analysis of failure times measured on dairy cows: Theoretical considerations in animal
breeding. J. Dairy. Sci. 69: 217-227.
Sorensen D (1996). Gibbs sampling in quantitative genetics. Intern Report, Danish Institute of Animal Science, Department of
Breeding and Genetics, Tjele.
Van Vleck LD (1980). Stayability evaluation as a categorical trait and by considering other traits. J. Dairy. Sci. 63: 1172-1180.

Genetics and Molecular Research 14 (4): 14956-14966 (2015) FUNPEC-RP www.funpecrp.com.br