|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
MEXICAN SIMMENTAL-BRAHMAN GENETIC CHARACTERIZATION, GENETIC
PARAMETERS AND GENETIC TRENDS
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
To Serafin and Micaela, my parents.
To Javy and Enry, my family.
To Eduardo, Silvia, Jorge, Raul, Ignacio, Joel, Gabriel, Horacio and Olga Lidia, my
And to M. A. Elzo, the researcher.
I wish to express my gratitude to the Mexican National Council of Sciences and
Technology (CONACyT), to the Mexican National Institute of Forestry, Agricultural and
Livestock Research (INIFAP), to the Florida Mexico Institute (FMI) and also to the
Tamaulipas Produce Foundation for their financial support. I also wish to thank the
Mexican Simmental-Simbrah Association for the data upon which this research is based.
The author wishes to express his sincere gratitude and appreciation to Dr.
Timothy A. Olson, chair of his supervisory committee and to the rest of the supervisory
committee, Dr. Ramon C. Littell, Dr. Don R. Sloan and Dr. David G. Riley, for their
friendly attitude and valuable advice throughout this study and the implementation of this
I greatly appreciated the friendship and encouragement offered by friends and
colleagues. I wish to give special thanks to Virginia Rada, Pedro and Imelda Garces,
Rey Acosta, Arturo Bocardo, Moises Montafio, Vicente Vega and Arturo Reyes.
I especially wish to express my thankfulness to Dr. Kenneth Gerhardt, Ines
Aviles, Nora Infante, Debra Anderson, Jane Luzar, Martha Hartman, Lisa Praharani and
Glenda Tucker for the unconditional support in the darkest moment of my life.
Gratitude is extended to the good people of Animal Sciences Department.
I would like to express my gratitude to Enriqueta Duarte and little Javy, because
they were the light of my life and helped me and encouraged me in finishing this proj ect.
And to God, for the miracle of the life and because He is taking care of me.
TABLE OF CONTENTS
ACKNOWLEDGMENT S .............. .................... iv
LI ST OF T ABLE S .........__.. ..... .___ .............._ vii..
LI ST OF FIGURE S ........._.___..... .__. .............._ viii..
1 INTRODUCTION ................. ...............1.......... ......
2 LITERATURE REVIEW .............. ...............5.....
Linear Models in Animal Breeding ................. ...............5............ ...
Genetic Computational Software........................ ...........2
Crossbreeding and Multibreeding Mixed Models .............. ...............27....
Cytoplasmic Effects ............._ ......... ...............29.....
3 GENETIC PARAMETERS INT SIMMENTAL-BRAHMAN HERDS UNDER
MEXICAN SUBTROPICAL CONDITIONS ................. ............... ......... ...33
Introducti on ................. ...............33.................
M material and M ethod s .................. ...............3.. 4......... ....
Data and Herd Management ................. ............. ............ ....... ........ 3
Estimation of Variance Components and Heritabilities ................. ................. 37
Results and Discussion .............. ...............41....
Group Genetic Effects ........................ ...............4
Variance Components and Heritabilities............... .............4
Im plications .............. ...............48....
Sum m ary ................. ...............48.......... ......
4 GENETIC PARAMETERS AND TRENDS FOR PREWEANING GROWTH
TRAITS INT THE MEXICAN SIMMENTAL POPULATION............... ...............5
Introducti on ................. ...............50.................
Material and Methods ................. ...............51........... ....
Data and Pedigree Files ................. ...............51................
Estimation of Genetic Parameters .............. ...............52....
Genetic Trends............... ...............55.
Results and Discussion .............. ...............55....
Phenotypic results ............... ......__ ...............55....
Estimation of Genetic Parameters .............. ...............58....
Genetic Trends............... ...............60.
Im plications .............. ...............62....
Summary ........._.__....... .__ ...............62....
5 MATERNAL LINE EFFECT INT PRODUCTIVE TRAITS INT MEXICAN
SIMMENTAL-SIMBRAH POPULATION ......._ ......... ___ ........._ ......64
Introducti on ................. ...............64.................
Material and Methods ................. ...............66........... ....
Data and pedigree files .............. ...............66....
Estimation of Genetic Parameters .............. ...............66....
Results and Discussion .............. .. ...............68...
Estimation of Genetic Parameters .............. ...............68....
Im plications .............. ...............74....
Sum m ary ............. ...... ._ ...............74....
6 GENERAL CONCLUSIONS ............. .....___ ...............76...
APPENDIX: GRAPHIC REPRESENTATION OF NON-ADDITIVE GENETIC
EFFECT ON CROSSBRED ANIMALS................ ...............80
LIST OF REFERENCES ............. ...... ._ ...............84....
BIOGRAPHICAL SKETCH .............. ...............94....
LIST OF TABLES
2-1 Wall clock time (in minutes) to solve three mixed model equations of different size
(Ml, M2, M3) and number of iterations until convergence, by different computing
softw are. ............. ...............23.....
3-1 Number of sires, dams and calves by herd and breed-group-of-sire 0I breed-group-
of-dam combination .............. ...............36....
3-2 Fixed environmental effects, fixed genetic group effects, random genetic effects
presented in the model, by herd and trait .............. ...............40....
3-3 Least square means ( A S.E.) for birth weight and weaning weight by breed group
of dam in herd 1 .............. ...............42....
3-4 Estimates of genetic variances and heritability ( A S.E.) for birth weight (BWT)
and weaning weight (WW205) in all herds ........................... ........._._ .....4
4-1 Least square means ( + S.E.) for birth weight and weaning weight by calf sex,
season and age of dam for Mexican Simmental breed ................. ............. .......56
4-2 Covariance components and genetic parameters for birth weight and weaning
weight in the Mexican Simmental population............... ...............5
5-1 Covariance components and genetic parameters and standard error for birth weight
in the Mexican Simmental-Zebu purebred and crossbred population......................70
5-2 Covariance components, genetic parameters and standard errors for birth weight
and Weaning Weight in purebred Mexican Simmental population .........................71
5-3 Covariance components, genetic parameters and SE for birth weight and weaning
weight in Mexican Simmental-Zebu crossbred population............... ...............7
LIST OF FIGURES
4-1 Distribution of Simmental herds in Mexico by state ................. .......................51
4-2 Birth weight by birth year of Mexican Simmental breed ................. ................ ..57
4-3 Weaning weight by birth year of Mexican Simmental breed ................. ...............57
4-4 Trends for birth weight calf additive direct (BWD) and dam additive maternal
(BWM) predicted genetic values in the Mexican Simmental population. ...............61
4-5 Trends for weaning weight calf additive direct (WWD) and dam additive maternal
(WWM) predicted genetic values in the Mexican Simmental population............_...61
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
MEXICAN SIMMENTAL-BRAHMAN GENETIC CHARACTERIZATION, GENETIC
PARAMETERS AND GENETIC TRENDS
Chair: T. A. Olson
Major Department: Animal Sciences
Covariance components for birth weight direct (BWTD), birth weight maternal
(BWTM), weaning weight direct (WW205D) and weaning weight maternal (WW205M)
were estimated in a Mexican Simmental-Simbrah population. In the first study, data
from three disconnected herds where the cattle were reared under subtropical conditions
were used to estimate genetic additive and nonadditive parameters. Estimates of
regression genetic group effects for BWT and WW205D showed an important effect of
nonadditive factors on each of these traits in all three herds. Thus, these nonadditive
effects should be considered in any genetic improvement program. The genetic
parameters that were estimated indicated more genetic variability for weaning weight
than for birth weight in these herds. A second study using purebred Simmental data from
the Mexican Simmental-Simbrah Association was conducted. Estimate of heritability for
birth weight additive direct was 0.40 while it was 0.12 for additive maternal. The genetic
correlation between additive direct and additive maternal effects was -0.63 for birth
weight. Weaning weight heritabilities were 0.33 for additive direct and 0.19 for additive
maternal. The additive direct-maternal correlation for weaning weight was -0.67. The
only across-trait nonzero genetic correlation was between birth weight direct and
weaning weight direct (0.73). Genetic parameters found in this study were almost
identical to those found in the American Simmental population, suggesting that the
Mexican Simmental population is a subset of the American Simmental population. A
third study was conducted using data from a purebred and crossbred Simmental-Simbrah
population to determine the effect of cytoplasmic line. Only small cytoplasmic effects
were found for BWT and WW205D ranging from 2 to 3%. The genetic correlations
between direct and maternal effects, between direct and cytoplasmic effects and between
maternal and cytoplasmic effects were 0.12, 0.29 and 0.98, respectively. Even though the
cytoplasmic effect was small, it may be useful in the selection of maternal lines to use in
specific genetic improvement programs such as cloning and embryo transfer within the
Mexican Simmental-Simbrah population.
Performance of livestock in tropical regions of the world is known to be inferior to
that of livestock in temperate zones. Attempts to improve cattle production in the tropics
have met with only limited success. Technological advances developed in industrialized
countries and applied to tropical animal production have not generally worked well. In
general breeds imported from temperate zones have not adapted well to the extreme
conditions encountered in many of the tropical climates (Pearson de Vaccaro, 1973;
Olson et al., 1990; Rios-Utrera et al., 1996; Vega-Murillo et al., 1996; Rosales-Alday et
Large amounts of bovine semen and live animals have been exported from the
United States to Mexico. Semen has been used to improve Mexican beef production, but
no evaluation has been made to quantify the genetic impact of introduction of new genes
to the Mexican beef population. The commercial relationship between the United States
and Mexico is important and Mexico has been introducing large quantities of germplasm
from the USA. The imported breeds with the largest numbers in the tropical regions of
Mexico are purebred Simmental and Simbrah. Cattlemen are using live sires and dams
and frozen semen from these breeds on the Zebu based population of the region.
However, crossbreeding between Simmental and Zebu has been done without a clear
strategy and performance of these animals under tropical and subtropical conditions is
Mexican tropical beef producers are trying to increase productivity by introducing
new breeds such as Angus, Charolais, Limousin, Simmental, Holstein and Brown Swiss
to the Zebu- based population (Rios-Utrera et al., 1996; Vega-Murillo et al., 1996). The
obj ective of Mexican cattlemen is to increase production while maintaining some Zebu
influence for adaptation. Beef production can be improved using appropriate sources of
germplasm and properly designed mating systems. Information is needed on expected
performance of available breeds and their crosses for economically important
characteristics, under varying levels of management, in different climatic conditions and
in different geographic areas (Falconer, 1989; Weigel, 2002). The breeding value of
imported animals is usually unknown. Perhaps some of the imported animals may have a
genetic evaluation in their country of origin, but these breeding values may be different
under different environment and production systems.
Another alternative for increasing production in the tropics is the genetic
improvement and wider distribution of the native breeds which are naturally adapted to
their surroundings (Alves Santiago, 1976). Unfortunately, little documented information
is available on the productive traits and behavior of these indigenous breeds.
Reproductive efficiency has been reported as poor (Randel, 1984; Galina and Russell,
1987; Galina and Arthur, 1989, 1991). Mexican cattle breeders have recognized the
necessity of having genetic evaluations of their livestock. Several factors, however, limit
their ability to do so at this time. One of these limiting factors is a lack of qualified
personnel; another is the absence of a proven genetic program. To solve these problems,
a genetic program was proposed with the participation of Mexican National Institutes of
Research, universities and cattlemen's associations (SAGAR, 1998).
In Mexico the bovine population has a truncated pyramidal structure with
commercial producers forming the base. The middle is formed by breeders who import
sires from others countries. The top part of this pyramid is missing because specific
breeding plans to include selection within the entire Mexican beef cattle population are
lacking (De los Santos et al., 2003). In order to solve this situation, the Mexican
government has implemented a breeding plan in which genetic evaluations are included.
This plan includes the participation of Mexican beef breeder associations, universities
and the National Institutes of Research in Forestry, Agriculture and Livestock Production
(INIFAP) (SAGAR, 1998).
Zebu breeds are widely used in the Mexican tropics for beef production. However,
declining values of feeder calves with Zebu characteristics have convinced cattlemen to
use alternative crossbreeding systems using Bos taurus sires (Angus, Charolais,
Limousin, Simmental, Holstein and Brown Swiss) on Zebu dams. Several breed
comparisons have been conducted involving crosses of Bos taurus sires and Bos indicus
cows under tropical Mexican conditions (Rios-Utrera et al, 1996; Vega-Murillo et al.,
1996; Quiroz-Valiente et al., 1994). These authors concluded that Bos taurus x Bos
indicus crossbred Fl animals were younger at first calving and had a higher weaning rate.
However, no genetic evaluation has been conducted in this region. Precise estimates of
genetic parameters are required to define genetic strategies to increase animal production
under Mexican conditions.
Another factor to consider is that the genetic evaluations to date have been made
using a one breed approach, considering only additive genetics effects. However, when
sires are intended to be used on cows of a different breed, another approach is needed.
This approach is called a multibreed analysis (Elzo and Bradford, 1985; Elzo and
Famula, 1985). Using this procedure, it is possible to evaluate not only additive genetic
values, but also non-additive effects as well. This approach allows one to determine
which sire has the best performance with a particular breed group of dams (Elzo, 1999).
Genetic improvement in the USA has been based on selection for economically
important traits such as birth weight, weaning weight, yearling weight and others (ASA,
2003). The Mexican database includes only birth weight and weaning weight.
Consequently, the series of studies in this dissertation are expected to achieve the
following obj ectives:
1) to estimate genetic parameters for birth weight (BWT) and 205 d-adjusted
weaning weight (WW205) for three disconnected herds under Mexican subtropical
conditions (Chapter 3),
2) to estimate additive direct and maternal genetic variances and covariances,
heritabilities, genetic correlations and genetic trends for birth weight and weaning weight
in the Mexican Simmental population (Chapter 4) and
3) to determine the contribution of cytoplasmic line effect on birth weight and
weaning weight (Chapter 5).
Some countries are importers and others exporters of animal products with distance
not being an issue anymore. For this reason the product that cattlemen are selling or
buying should be of the best quality possible. Genetic improvement of cattle has become
a global venture. Producers routinely use semen from foreign sires and breeding
companies acquire genetics from a variety of countries. The international exchange of
cattle, semen and embryos necessitates a methodology to compare animals that differ in
housing conditions, feeding programs and genetic composition. Years ago, genetic
differences between strains of cattle in different countries were large and to identify
which breed, line or animal was superior, a simple comparison of phenotypic records was
sufficient. Over time countries that were once importers of genetic material became
competitors, with export aspirations of their own. Genetic differences between strains
within various countries are now smaller and advanced statistical procedures are needed
to compare animals fairly and accurately (Weigel, 2002). Currently efforts exist to
analyze information generated from beef and dairy cattle (Canavesi et al., 2002;
Emmerlin et al., 2002; Lidauer et al., 2002) from several countries simultaneously.
Linear Models in Animal Breeding
Prediction of breeding values constitutes an integral part of most breeding
programs for genetic improvement. Crucial to the accurate prediction of breeding values
is the availability of records. In a population, data available at the initial stages are
usually from individual animals which may or may not be related. Data collected later
may include information on offspring and other relatives. Initially the prediction of
breeding values may be based on records of individuals and a few relatives. To analyze
production records and get a genetic evaluation, it is necessary to model the data.
Modeling is based upon the assumption that every phenotypic observation of an animal is
determined by environmental and genetic factors and may be defined by the following
Phenotypic observation = environmental effects + genetic effects + residual effects
y?1 = CRi + gi + eij
y, is the record j of the animal ith,
CRi includes the identifiable non-random (fixed) environmental effects such as herd
management, year of birth, sex of the ith animal,
gi is the sum of the additive (ga), dominance (gd) and epistatic (ge) genetic values of
the genotype of the ith animal and
e, is the sum of random environmental effects affecting animal i.
The additive genetic value in the g term above represents the average additive
effects of genes an individual receives from both parents and is termed the breeding
value. Each parent contributes a sample half of its genes to its progeny. The average
effect of the sample half of genes which a parent passes to its progeny is termed the
transmitting ability of the parent and corresponds to one-half of its additive genetic value
(V~ga). The breeding value of the progeny, therefore, is the sum of the transmitting
abilities of both parents. Since the additive genetic value is a function of the genes
transmitted from parents to progeny, it is the only component that can be selected for and
therefore is the main component of interest. In most cases, dominance and epistasis,
which represent intralocus and interlocus interactions respectively, are assumed to be of
little significance (Falconer, 1989; Mrode, 1996) and are included in the e, term of the
y?1 = CRi + gai + e ,
with egi including the sum of the random environmental effects, dominance and epistatic
genetic values. This equation constitutes the linear model usually employed in most
problems of breeding value prediction in animal breeding (Henderson, 1975b; Fries,
1998). Usually it is assumed that y, follows a multivariate normal distribution (MVN)
(Searle, 1971), implying that traits are determined by an infinite number of additive genes
each with infinitesimal effects at unlinked loci; the so-called infinitesimal model
(Falconer, 1989; Weigel, 2002). Also it is assumed that var(gai) and var(e gj) are known,
or at least, their proportionality is known and that there is no correlation between gai and
e*1: (cov(gai, e* ) = 0 and no correlation among mates (cov(e 4i e ik )= 0. Also CL
represents the mean performance of animals in the same management group
(contemporary group). The mean of animals reared under the same management system
and of the same age and sex is assumed to be known. This reduces the problem of
predicting breeding values to adjusting phenotypic observations for identifiable non-
random environmental effects and appropriately weighting the records of animals and
their available relatives.
From the previous model one should define ai as the breeding value of animal i,
ai = gai = Ias + 1/ ad + mi
where as and ad are the breeding values of the sire and dam respectively and mi is the
deviation of the breeding value of animal i from average breeding value for both parents,
that is, Mendelian sampling (Mrode, 1996). The sampling nature of inheritance implies
that each parent passes only a sample half of its genes to its progeny. Therefore, genetic
variation exists between offspring of the same parents since all offspring do not receive
exactly the same genes. Mendelian sampling could be regarded as the deviation of
average effects of additive genes an individual receives from both parents from the
average effects of genes common to all offspring of the same parents (Elzo, 1996).
Accurate prediction of breeding value constitutes an important component of any
breeding program since genetic improvement through selection depends on correctly
identifying individuals with the highest true breeding values. The method employed for
the prediction of breeding value depends on the type and amount of information available
on candidates for selection.
Mixed Model Equations
Henderson (1953) developed a methodology called best linear unbiased prediction
(BLUP) by which fixed effects and breeding values can be simultaneously estimated.
The BLUP acronym stands for:
Best means it maximizes the correlation between true (a) and predicted breeding value
(A) or minimizes prediction error variance (PEV), (var(a A))
Linear predictors are linear functions of observations
Unbiased estimation of realized values for a random variable such as animal breeding
values and of estimable function of fixed effects are unbiased E[1^v] = E[w] and
Prediction involves prediction of true breeding values.
BLUP has found widespread usage in genetic evaluations of domestic animals
because of its desirable statistical properties. Its usage has been enhanced by the steady
increase in computing power and has evolved in terms of its application from simple
models such as the sire model in its early years to more complex models such as the
animal, maternal and multivariate models in recent years.
Matrix algebra can be used to describe the power of Henderson' s mixed model
equations. These equations have been and will be used extensively by animal breeders
and will continue in the future to be the basis for linear mixed model analyses. Let y be a
vector of observations of order n. The general linear model for the observations can be
p is a vector of fixed effects,
X is the matrix that associates the fixed effects with the observation in y and
a is a vector of random effects.
E~y] = X P
V(y) = V(e) = V of order n x n.
The generalized least squares solution for P is obtained from the generalized least
(X'V- X) S = X'V- y
The problem is that V-1 must be calculated. Computational time to obtain inverses
of matrices such as V-1 is proportional to n3. Thus, with the number of observations to
analyze, which currently are of the order of thousands and some countries have millions
of records, obtaining the inverse is impossible. If X'V- X can be calculated, iterative
methods can be used to obtain solutions to the equations, but obtaining V-1 for all except
trivial problems may still be impossible. The generalized least squares equations do
define solutions for the fixed effects P Such solutions denoted as P yield X P are best
linear unbiased estimators (BLUE) of X P But what animal breeders need are predictors
of random effects such as breeding values. These are included in E as in the mixed linear
a = Zu + e
u is the vector of random effects to be predicted (e.g., breeding values),
Z is the matrix associating effects in u with the animal record (y) and
e is the vector of random effects uncorrelated with u.
Predictors of elements of u which have the properties of BLUP, best linear
unbiased predictors, can be obtained as E[uly] with the restriction that E[G] = E[u]. One
way to obtain G is
G = GZ'V- (y X P)
where GZ' = cov(u,y'). The corresponding selection index predictor (also called linear
u = GZ'V- (y X P)
where p is the solution of fixed effects, is assumed to be known exactly instead of using
p the BLUE of f To adjust for Eixed effects to obtain BLUP, not only must V-1 be
calculated, but also p must be obtained from solving the generalized least squares
The problem is to obtain P and (1 without inverting V. Even with the assumption
that good estimates of variance components represented in V are available, the problem is
how to obtain V- Henderson (1953) gave a solution; it was intuitive and did not involve
calculation of V- He very likely tried to Eind efficient ways to do that, but was
unsuccessful. What he did do was this: if elements of u are assumed to be fixed effects,
then for the linear model
y = X p + Zu +e
the weighted least squares equations to solve when assuming both P and u are fixed are
X'R 'X X'R-'Z P X'R 'y
Z'R 'X Z'R-'Zj ^ -Z'R 'y
where R = V(e), which usually has a simple structure: diagonal and homogeneous for
single trait or block diagonal for multiple traits, so that R^1 is easily obtained.
Henderson's first formulation involved single trait equations multiplied through by 02,
the homogeneous variance of R = 102. In that case, the equations are called ordinary least
TX'Xzl X'ZlL TX'yl
ZX Z'Z ^I h Z'y
In the general formulation, Henderson added G-1 where G = V(u) to the Z'R^ Z
block of the least squares equations:
rX'R 'X X'R 'Z j rX'R' y
LZ'R 'X Z'R 'Z + G I ^ Z'R 'y
Henderson proved several years later that the solutions to these "modified" least
squares equations were, in fact, BLUE for P and BLUP for u. Because R^1 is usually
easily obtained and in many cases the structure of G is such that G-1 is also easily
obtained, then simple iteration can be used to obtain P and (1 from the mixed model
equations, known to most animal breeders and statisticians as Henderson's mixed model
equations. The proofs that solutions from mixed model equations are the BLUE of P
and BLUP of u are in Henderson et al. (1959).
Another contribution of Henderson to animal breeding is a method to obtain G- .
Wright (1922) worked out a measure of this genetic relationship between pairs of animals
called the numerator relationship. A numerator relationship matrix contains the
relationships among any number of animals. Having a fraction of genes in common is
like a fractional replication of genetic effects for related animals. Animal breeders
typically want to make use of this partial genetic replication through Ao2 to obtain more
accurate predictions of breeding values or to predict breeding values for animals without
records. In models typically used in animal breeding the number of animals may range
from thousands to millions. However, to apply Henderson's mixed model equations, G-1
rather than G is needed. Thus, for G = Ao2, what is needed is A- G- = A- 02,. To solve
this problem Henderson (1975a) published a simple set of rules to obtain elements of A-l
without having to calculate A. With those rules a list of animals with each animal's sire
and dam (pedigree file) is needed. The A-l matrix computed by Henderson's rules is of
the order of the number of unique animals, including sires and dams, some of which may
not have records. Thus, equations need to be included for all animals having records.
Usually the least squares equations will involve only the model for the available records.
In 1976 Henderson published a procedure to include animals without records in the
genetic evaluation, augmenting equations to the mixed model (Henderson, 1976a, b, c ).
These equations are tied to the equations for animals with records with the G- = A- 02,
part of the mixed model equations. Solutions for animals without records are basically
selection index predictions using relationships to relatives with records and their
predicted breeding values. This method is particularly valuable for multiple trait models
and for models with both direct and maternal effects when all animals do not have
measurements for all traits, or not all cows become mothers of progeny with records
(Henderson, 1975b, 1976d).
The sum of Henderson' s discoveries of 1) the modification of least squares
equations to form mixed model equations with solutions that have BLUE and BLUP
properties, 2) the easy calculation of A-l and 3) the augmentation of the mixed model
equations to include equations for animals without records (for all or some traits) has
provided animal breeders worldwide with the best possible tools for genetic selection
(Van Vleck, 1991).
The obj ective of the animal model is to predict the breeding value (BV) of animals
based on their own records and (or) records of their relatives. The animal model has the
1) animals belong to a single population,
2) animals may have one or more records and covariances among records that are due
only to genetic factors,
3) there is either no selection in the population or,
4) if selection occurred based on records, the selection was within fixed effects and
5) if selection occurred based on the BV of animals, the relationship matrix was
The animal model in matrix notation is
y =Xb +Zu +e
E[y] = Xb
And because of assumptions (i) and (ii), we can say that
du l G 0;-[
Var(y) = ZGZ' + R
= ZAZ' a:, + Io2
y = vector of animal records,
b = vector of unknown fixed effects,
u = vector of unknown random BVs belonging to the animals making records,
e = vector of unknown random residual effects,
X = known incidence matrix relating records to fixed effects in vector b,
Z = known incidence matrix relating records to BVs in vector u.
Then the mixed model equations (MME) for the animal model, after multiplying
both sides by 0; are
LZ'X Z'Z+A-'all11 uZ'yl
where Al is the inverse of the matrix of additive relationships among the animals with
The matrix A is the relationship matrix and it indicates the additive genetic
relationship among individuals. This matrix is symmetric and its diagonal element for
animal i (aii) is equal to 1 + Fi, where Fi is the inbreeding coefficient of animal i
(Falconer, 1989). The off-diagonal element, aij, equals the numerator of the coefficient of
relationship between animal i and j. When multiplied by the additive genetic variance
(a o), Aos is the covariance among breeding values. Thus, if u, is the breeding value for
Var(ui) = aii aof = (1 + Fi) o:
The matrix A can be computed using path coefficients, but when there is a large
amount of animals it is time consuming to do so. Henderson (1976a) proposed a
recursive method that is suitable for computation and is the method currently used.
Maternal Trait Models
The phenotypic expression of some traits in progeny, such a weaning weight in
beef cattle, is influenced by the performance of the dam. Thus, the dam contributes to the
performance of the progeny through genes that she passes to her progeny and also
through her ability to provide a suitable environment (for example milk production).
Traits such as birth and weaning weights in beef cattle fall into this category and are
termed maternally influenced traits. The ability of the dam to provide a suitable
environment for expression of such traits in her progeny is partly genetic and partly
environmental. Like the genetic component of an individual, the maternal genetic
component can be partitioned into additive, dominance and epistatic effects (Willham,
1972). The environmental component may be portioned into permanent and temporary
effects. It is the maternal additive genetic component of the dam that is passed on to her
offspring, but it is expressed only when the female offspring have progeny of their own.
In the usual mixed linear model for maternally influenced traits, the phenotype is
partitioned into the following:
1. Additive genetic effects from the sire and the dam, usually termed direct genetic
2. Additive genetic ability of the dam to provide a suitable environment, usually
termed indirect or maternal genetic effects.
3. Permanent environmental effects, which include permanent environmental
influences on the dam's mothering ability and maternal non-additive genetic effects of
the dam and
4. Other random environmental effects, termed residual effects.
The model for maternally influenced traits, in matrix notation, is
y =Xb +Zu +Wm +Spe +e
y = vector of observations
b = vector of fixed effects
u = vector of random animal effects
m = vector of random maternal (indirect) genetic effects
pe = vector of permanent environment effects
e = vector of random residual effects
X, Z, W and S are incidence matrices relating records to fixed animal, maternal
genetic and permanent environmental effects respectively.
It is assumed that
u g,,A gl2A 0 0
m K21A g22A 0 0
pe 0 0 Icr, e
gol is the additive genetic variance for direct effects
g22 is the additive genetic variance for maternal effects
gl2 is the additive genetic covariance between direct and maternal effects
ap is the variance due to permanent environmental effects
a; is the residual error variance
The variance of y is
rg, A gA lZ'
var(y)= [Z W1 Sla' S'+a"
Lg~zA g,,Alw W' p
The best linear unbiased estimator (BLUE) of estimable functions of b and the
BLUP of u, m and pe are obtained by solving the following mixed-model equations
X'X X'Z X'W X'S I b Xy
Z'X Z'Z +A-'a, Z'W+A-'a3 Z'S u Z'y
W' X W' Z +A-'a W' W +A 'a, W' S W'y
S' X S' Z S'W S' S+Ia4l -S' Ly
Application of a sire model implies that only sires are evaluated, using progeny
records. Most early applications of BLUP for the prediction of breeding values,
especially in dairy cattle, were based on this model. The main advantage with the sire
model is the number of equations is reduced compared with an animal model since only
sires are evaluated. However, with a sire model, the genetic merit of the mate (dam of
progeny) is not accounted for. It is assumed all mates are of similar genetic merit and
this can result in biases in predicted breeding values if there is preferential mating (Elzo,
1996; Mrode, 1996).
The model in matrix notation is:
y =Xb +Zs +e
All terms were defined earlier and s is the vector of random sire effects and Z
relates records to sires and
Var(s) = A og
Var(y) = ZAZ' 0,2 + R
Where A is the numerator relationship matrix for sires,
The MME are exactly the same as animal model except that
a = 0,2 /a, = (4-h2)/h2
Sire-Maternal Grand Sire Model (SMGSM)
An approach to estimation of maternal effects avoiding the bias given by the non-
maternal genetic effects is the use of the Sire-Maternal Grand Sire Model. The model
rests on simplifying assumptions made with respect to the data. The obj ectives can be to
reduce computations or to generate a better behaved set of mixed model equations. The
assumptions for this model are:
1) Parents have no records and
2) Dams are related only through their sires.
These simplifying assumptions reduce an animal model to a sire-maternal grand
sire model (Elzo, 1996).
Connectedness can be defined as a measure of the relationships between herds or
contemporary groups as they affect the accuracy of comparing the genetic values of
animals from one herd or group to another. When there is higher connectedness, there
are more accurate comparisons of expected breeding values (EBVs) across groups or
herds. Therefore, it is important to measure the degree of connectedness and, if
necessary, bring it to a level that allows comparison of EBVs with reasonable accuracy
(Mathur et al., 1998).
Genetic evaluations are used to compare individuals based on their relatives'
breeding values. When comparing individuals from different countries, or other definable
sub-populations, predicted differences between individuals include predictions of average
genetic differences between the sub-populations. There may also be environmental
differences affecting performance in each sub-population and the evaluation model must
therefore partition the observed differences in performance into average environmental
and genetic effects for each sub-population.
Genetic connections between sub-populations are required in order to separate
genetic and environmental portions of performance differences between the sub-
populations. The term connectedness refers to the number and quality of these genetic
connections. Kennedy and Trus (1993) reviewed the concepts of connectedness and
implications for genetic evaluation, describing situations where sub-population genetic
differences would not be estimable or would have high variance of prediction error
(PEV). The authors gave special attention to models with genetic groups, since
connectedness is particularly important with these models. Hanocq et al. (1996) also
found important effects of connectedness on PEV of genetic differences between sub-
populations. Connectedness varies among countries and populations (e.g., Banos et al.,
1991) and may be of practical concern for PEV of genetic differences between countries
in international evaluations.
Connectedness could also affect the accuracy of covariance component estimates.
Schaeffer (1975) suggested only well-connected data subsets should be used to estimate
genetic variances. Eccleston (1978) rebutted with a proof that all data should always be
used. Bans and Sigurdsson (1996), however, showed that with international data
undergoing selection within and across countries, the use of all data resulted in biased
estimates of genetic correlations between countries. Use of a well-connected data subset
gave unbiased estimates. The importance of connectedness clearly depends on the
evaluation model being used and characteristics of data being evaluated. Connectedness
is most important when genetic and environmental factors in the model are confounded,
unless there are genetic connections between the sub-populations and when data are
undergoing selection forces that may not be fully accounted for by the model.
When animals are reared under different environmental conditions, the accuracy of
sire comparison depends upon the degree of connectedness among those groups.
Kennedy and Trus (1993) argued that the most appropriate measure of connectedness is
the average PEV of differences in EBV between animals in different management units.
However, computing this statistic is extremely time consuming and not feasible for
routine application (Roso et al., 2002). Thus, Kennedy and Trus (1993) proposed the use
of variance of estimated differences between management unit effects, which were highly
correlated with the PEV of the differences between EBV in their simulation study.
Mathur et al. (1998) argued that the PEV of difference between herd effects can be used
to measure the degree of connectedness between two herds. They proposed the
connectedness rating (CR), defined as the correlation between estimated effects of two
management units. This CR is less dependent on size and structure of management units.
Fries (1998) proposed the use of a number of direct genetic links (GLt) between different
management units due to common sires and dams as a method for measuring degree of
connectedness among contemporary groups. Roso et al. (2002) described the comparison
between connecting rating and genetic links for measuring the degree of connectedness
among contemporary groups of station-tested bulls. They concluded CR is more
dependent on the size of the CG than GLt and the use of CR as a connectedness measure
will cause larger contemporary groups (CG) to be favored. GLt, unlike CR, allows
differentiation between completely disconnected CG from connected ones. GLt is
computationally much less demanding than CR and can be routinely used with the aim of
increasing the level of connectedness, obtaining more accurate comparisons of EBVs
Genetic Computational Software
The complexity of models used or considered for use in genetic evaluation is
increasing. Examples of new models are test-day models in dairy cattle (Meyer and Hill,
1997; Gengler et al., 1999; Meyer, 2002), growth models in beef cattle or random
regression models (Schaeffer, 2004) and models with dominance and/or quantitative trait
loci (QTL) effects (Broman, 2002; Szyda et al., 2002). These models are usually linear
but analyses of some traits may require nonlinear models which are usually more
complicated to write and test. More complicated models may require larger data sets. In
the future one may expect new types of models that will be used to analyze even larger
data sets (Misztal, 1999).
Efforts have been made to make software more efficient. Lidauer and Stranden,
(1999) compared two procedures, MiX99 and DMUIOD and it is clear there is genetic
software that is more efficient than others. Time required to Einish an analysis depends
on the number of records and complexity of the model (Table 2-1). Other new
alternatives for data analyses include use of parallel computers and scalability to increase
memory and reduce computational time (Larsen and Madsen, 1999; Madsen and Larsen,
Table 2-1. Wall clock time (in minutes) to solve three mixed model equations of different
size (Ml, M2, M3) and number of iterations until convergence, by different
Computing software MiX99 DMUIOD
Ml M2 M3 Ml M2 M3
Time to prepare data for solver 2 36 62 2 66 189
Time for solving until convergence 7 161 490 55 1,116 10,152
Number of iterations 212 149 167 438 305 380
Ml: 240,000 records and 38,256 animals, single trait.
M2: 6.7 million records and 1,343,337 animals, single trait.
M3: 8.4 million records and 1,343,337 animals, three traits.
Lidauer and Stranden, 1999.
To analyze genetic information there are several computational packages that will
This is a general mixed models analysis program with emphasis on the estimation
of variance components under a range of variance structures. It was first written in 1996
and is under continuing enhancement with particular attention to agricultural
experiments. The two central ideas incorporated in the ASREML program are use of the
Average Information (AI) algorithm for obtaining the Restricted Maximum Likelihood
(REML) update to the variance parameters and use of sparse matrix methods (Gilmour et
al., 1995). The Average Information algorithm is a quadratic convergence method
similar to Fisher Scoring except that it uses an information matrix which is approximately
the average of the observed and expected information matrices. These both contain a
trace term which takes considerable time to calculate. But when the average is formed,
the trace term is canceled. The average information matrix is calculated avoiding the
calculation of the trace term. The average information algorithm is supplemented by
Expectation-Maximization (EM) steps when the matrix estimated by AI is non-positive
definite. The second idea is based upon the assumption that not all elements of the
inverse of the coefficient matrix are required. The equations are solved in an order that
retains a high level of sparsity (Gilmour et al., 2002).
DFREML and MTDFREML
Derivative Free Restricted Maximum Likelihood and Multitrait Derivative Free
Restricted Maximum Likelihood software have been written with the analysis of data
from animal breeding in mind. They deal with the estimation of covariance components
and the resulting genetic parameters, e.g. heritabilities and genetic correlations. In
calculating the likelihood it is assumed that data have a multivariate normal distribution.
The model of analysis fitted throughout is the Animal Model and these programs allow
one or two additional random effects to be fitted. A total of 10 different models of
analysis are accommodated, depending on the variances and covariances to be estimated
and the assumptions in the model.
Maximum Likelihood Estimation: (ML) is conceptually straightforward. For a
given model of analyses, one set of parameters to be estimated with an assumed
distribution of the data, we can write the likelihood function L. The ML estimates for a
specific data set are the numerical values of the parameter for which L attains its
maximum. The likelihood function is commonly assumed to be proportional to the
probability density function of the distribution invoked i.e., the likelihood concept is
related to probability calculations. After the parameters are estimated, data prediction is
made to make inferences about the parameters which have given rise to observed data. In
practice, the log of the likelihood function, often referred to as the support function is
maximized (Meyer, 1997).
The scope of ML estimation for estimation of variance components has been
reviewed by Elzo (1996). While ML estimators have 'built-in' optimal statistical
properties such as consistency, efficiency and asymptotic normality and, under certain
conditions, can account for non-random sampling of data, there are two maj or conceptual
drawbacks. The first is the distribution of the data, usually a multivariate normal
distribution in the estimation of variance components, is assumed to be known. There are
some suggestions that the ML estimator assuming normality may be appropriate even if
this does not hold. Secondly, ML estimators are biased by the Eixed effects and the Eixed
effects in the model are treated as if they were known. This bias can be removed by
considering only the part of the likelihood function which is independent of the Eixed
effects (Meyer, 1997). This procedure is generally referred to as Restricted Maximum
Likelihood (Searle, 1971).
In animal breeding, interest in (RE)ML estimation of (co)variance components has
been stimulated because it yields estimates less affected by selection bias than analysis of
variance type methods (Henderson, 1975b). More recently, attention has focused on its
desirable statistical properties as well as its flexibility. This allows for non-standard
designs in estimating genetic parameters, i.e., the combination of information from
different types of relatives (Henderson, 1976b). Developments in the application of
REML to analyze animal breeding data have followed closely those in the genetic
evaluation of animals by Best Linear Unbiased Prediction (BLUP) (Henderson, 1976a).
In particular, the introduction and increasingly widespread use of the animal model has
furthered interest in and use of REML. Except for simple and balanced designs, REML
estimation of variance components requires the numerical solution of a constrained, non-
linear optimization problem. Iterative procedures to locate the minimum or maximum of
a function are generally classified according to the amount of information from
derivatives of the function which is utilized (Meyer, 1997).
Early applications of REML under a sire model generally used algorithms requiring
first or second derivatives (or their expected values) of the log likelihood function, i.e.,
the EM (Expectati on-Maximization) algorithm and Fisher Scoring Method. Though
relatively fast to converge, these were computationally demanding. EM is known to be
slow to converge. They require the inverse of a matrix of size equal to the number of
random effects times the number of traits analyzed in each round of iteration. This was
true unless some special features of the data structure could be exploited. Alternatively,
the optimum of a function can be located without knowledge of its derivatives. Methods
range from direct search procedures, which rely on mere comparison of function values,
to procedures approximating first or even second derivatives using numerical techniques.
Use of a derivative-free (DF)REML algorithm was first considered by Graser et al.
(1987) for univariate analyses under an animal model. While they considered a model
fitting animals' additive genetic effects as the only random effect, the approach is suitable
for models including additional random effects and multivariate analyses (Meyer, 1989).
Crossbreeding and Multibreeding Mixed Models
Livestock in the tropical regions are known to be generally less productive than
those in more temperate zones (Galina and Russell, 1987). With approximately one-third
of the world' s cattle population situated in the tropics (FAO, 2003), this source of animal
protein must be fully exploited to meet the needs of the rapidly growing human
population. Attempts to improve cattle production in the tropics have met with limited
success. Technological advances developed in industrialized countries and applied to
tropical animal production generally have not worked well under local conditions.
Breeds imported from temperate zones have not adapted well to the extreme conditions
encountered in many tropical climates (Pearson de Vaccaro, 1973). Therefore, much
research effort has concentrated on development of crosses between local Bos indicus
cattle and Bos taunts breeds that have shown better adaptability to the tropics (Galina and
Russell, 1987; Galina and Arthur, 1989, 1991). Another effort has been to identify Bos
taunts breeds that have been adapted to the tropics. The obj ective is to use these breeds
in developing countries (Hammond et al., 1996; Elzo et al., 1998a, 2001; Pefia, 1998).
Elzo (1999) defined a multibreed population as one composed of straighbred and
crossbred animals that interbreed. Multibreed populations can be classified as either
complete or incomplete. Complete multibreed populations are those whose mating
patterns follow a diallel design. Incomplete multibreed populations are those having
different numbers and/or kinds of mating groups of sire and dams. Several approaches
have utilized in evaluation of multibreed populations.
The Multibreed Evaluation Procedure developed by Elzo and Famula (1985) allows
purebred and crossbred animals to be evaluated simultaneously. This approach allows
purebred and crossbred sires to be evaluated for additive, nonadditive and total genetic
effects. The total genetic value of an animal (TEPD) is defined as the sum of an additive
expected progeny difference (EPD), plus a non-additive EPD. The Additive EPD is that
portion of a multibreed evaluation equivalent to a within-breed EPD. The non-additive
EPD measures the combining ability of a sire when mated to dams of a particular breed
composition (Manrique, 1992; Elzo, 1999). This approach has been used to analyze
several crossbred populations: the Angus-Brahman herd of the University of Florida for
preweaning growth traits (Elzo and Wakeman, 1998) and carcass traits (Elzo et al.,
1998b); the Romosinuano-Brahman herd of Turipana, Colombia for pre and postweaning
growth traits (Elzo et al., 1998a) and the Sanmartinero-Brahman herd of La Libertad,
Colombia for pre and postweaning growth traits (Elzo et al., 2001). In the multibreed
population of dairy cows of Thailand, this approach has been used to estimate multibreed
genetic parameters and genetic values for first lactation 305-d milk, fat yield and fat
percentage. It has also been utilized in the estimation of lactation curves and prediction of
daily and accumulated milk yield (Koonawootrittriron et al., 2001, 2002a, b).
The current approach used to compare bulls across breeds consists of adding a
correction factor to the difference between the intrabreed EPDs of two bulls. These
correction factors are based on differences in performance of progeny of bulls of various
breeds when mated to Angus and Hereford females. These correction factors have been
computed at the U. S. Meat Animal Research Center (USMARC) in Clay Center,
Nebraska (Barkhouse et al., 1995; Gosey, 1997; Van Vleck and Cundiff, 2004). This
type of comparison may not be appropriate because it ignores differences between
environments and assumes that all animals in the analysis are of uniform genetic
composition (Nufiez-Dominguez et al., 1995).
In eukaryotic cells there are two kinds of DNA; nuclear DNA and mitochondrial
DNA. The inventory of nuclear and mitochondrial-coded proteins required to assemble a
functional mitochondrion shows clearly that nuclear and mitochondrial genomes interact
in at least two ways. First, both contribute to mitochondrial protein function. Secondly,
affect the synthesis and assembly of mitochondrial proteins. The first type of interaction
is important for the regulation of oxidative energy production. Isoforms of the nuclear-
coded subunits of cytochrome c oxidase affect the catalytic functions of its
mitochondrially-coded subunits. These isoforms are differentially regulated by
environmental and developmental signals and allow tissues to adjust their energy
production to different energy demands. The second type of interaction, biosynthesis of
mitochondrial proteins, requires the bidirectional flow of information between the
nucleus and the mitochondrion. Communication from the nucleus to the mitochondrion
makes use of proteins that are translated in the cytosol and imported by the
mitochondrion. Communication from the mitochondrion to the nucleus involves
metabolic signals and one or more signal transduction pathways that function across the
inner mitochondrial membrane. Just as the nuclear genome affects the expression of
mitochondrial genes, the mitochondrial genome can affect the expression of nuclear
genes for mitochondrial and other proteins (Poyton and McEwen, 1996).
Cattle breeders have long believed in special attributes of certain maternal lineages,
often referred to as cow families. For example, when they are buying a bull or dam, they
request not only for the genetic evaluation, but also for the maternal lineage and they
prefer animals that come from outstanding females. A possible cause could be correlated
environmental effects within lineages; perhaps cows born into outstanding lineages
receive preferential treatment in proportion to the perceived quality of the family (Gibson
et al., 1997).
There are three genetic explanations for maternal effects: maternal additive
inheritance, cytoplasmic inheritance and maternally-biased parental imprinting. Parental
imprinting occurs at some loci, some genes function properly only when they are donated
by the father, while others function properly only when they are contributed by the
mother (Surani et al., 1984; Hoffman, 1991).
The mitochondria (mt) provide a possible mechanism of cytoplasmic inheritance,
being inherited exclusively though the maternal lineage (Hutchison et al., 1974).
Arguments in favor of mitochondrial contributions to variation include: 1) the
mitochondria are central to cellular function, particularly energy metabolism; 2) there are
probably several thousand copies of mtDNA in an average cell and only two copies of
nuclear DNA; 3) mtDNA may code for up to 10% of gene products expressed in
mitochondria, though it may be much less given the lack of knowledge about the number
of enzymes and structural products involved in mitochondrial function; and 4) mtDNA
has about 10 fold higher mutation rate than nuclear DNA (Gibson et al., 1997).
Environmental effects for some economically important traits in beef cattle include
the maternal environment provided by the dam. A maternal effect is any influence, other
than the contribution resulting from nuclear genes that the dam has on the phenotype of
her progeny (Rohrer et al., 1994). Since mitochondria play an important role in
intracellular protein and energy metabolism, the assumption that cytoplasmic effects
could be involved in productive traits is reasonable (Bell, 1985).
Differences among maternal cytoplasmic lines in dairy cattle accounted for 2% of
the total variation for milk yield and 3.5% of total variation for fat percentage. They
were a significant source of variation for open days (Tess et al., 1987). Huizinga et al.
(1986) found cytoplasmic variation for milk yield was 5.6%, fat percentage was 4.8%,
protein percentage was 6.2%, kilograms of fat plus protein was 10. 1% and milk return
was 2.5% of the total variation.
Cytoplasmic effects in beef cattle accounted for 2% for birth weight, 5% for
average daily gain and 5% for weaning weight adjusted to 205 days in Hereford calves
(Tess et al., 1987). On the other hand, Brown et al. (1988) reported that variability in
mitochondrial respiratory activity had little association with weaning weight and yearling
growth in beef cattle. Cytoplasmic effects were not detected for birth weight, weaning
weight and post weaning average daily gain in Brangus cattle (Rohrer et al., 1994). In
sheep, no cytoplasmic effects were found for birth weight, weaning weight, fleece weight
and number of lambs born (Hanford et al., 2003; Van Vleck et al., 2003).
Under tropical conditions, Bos indicus x Bos taurus crosses have been used to
increase production traits (Elzo et al., 2001; Olson et al., 1990; Rosales-Alday et al.,
2004b). Roberson et al. (1986) found differences between reciprocal crosses for birth
traits. A difference of more than 7 kg for birth weight has been observed between
reciprocal cross embryos regardless of breed type of surrogate dam (Baker et al., 1990;
Thallman et al., 1992). Cytoplasmic inheritance could be responsible for some of these
Effects of mitochondrial DNA have been reviewed and recently estimated by
embryo transfer or by statistical procedures (Bell et al., 1985; Huizinga et al., 1986; Tess
et al., 1987). Bell et al. (1985) reported that cytoplasmic effects accounted for 2, 1.8, 1.8
and 3.5 % of the total variance for milk yield, milk fat, fat corrected milk and fat
percentage, respectively. Cytoplasmic effects were significant even after adjustment for
sire, maternal grandsire and dam's production. Northcutt et al. (1991) analyzed birth and
weaning weight for 847 and 427 synthetic beef cattle, using Jersey, Angus and
Simmental sires on crossbred cows, in three lines (small, medium and large size) divided
between two herds, representing up to 69 lineages and tracing back to foundation cows in
1977. Use of the animal model with lineage as a random effect gave estimates of lineage
contribution to total variance in different lines and herds from zero to 5%, with 8 of 12
estimates being zero. The authors concluded that there was no evidence for cytoplasmic
inheritance in those herds.
Important cytoplasmic genetic effects of beef cattle could have large effects on
breeding programs. Identification of superior cytoplasmic lines would increase the value
of some cows for use in embryo transfer. Similarly, breeds with superior cytoplasm
would become more valuable as maternal lines in crossbreeding systems or as foundation
females for composite breeds. Large cytoplasmic genetic effects would increase the
value of maternal relatives' performance (Tess et al., 1987).
When mitochondrial effects are ignored, the estimation of heritability and
permanent environmental variances and additive genetic variance were overestimated
(Boettcher et al., 1996). Southwood et al. (1989) obtained similar results and noted the
additional confounding of additive direct and cytoplasmic variance with the additive
GENETIC PARAMETERS IN SIMMENTAL-BRAHMAN HERDS UNDER
MEXICAN SUBTROPICAL CONDITIONS
Tamaulipas is a Mexican border state, located south of Texas. The commercial
relationship between these two states is important for Mexico and the USA and includes
livestock products such as live animals from Tamaulipas to Texas and bovine germplasm
from Texas to Tamaulipas. This germplasm has been used according to specific Mexican
cattlemen criteria. These criteria have not included predicted genetic values that could
increase the productivity of Mexican herds for economically important traits and would
increase quality of cattle sent from Mexico to the USA (De los Santos et al., 2003).
The imported breeds with the largest numbers in Tamaulipas are Simmental (S) and
Simbrah (SB). Males and females of these imported breeds have been crossed with
Mexican Brahman (B) cattle. Crossbreeding has been done without a clear strategy and
genetic ability of the resulting crossbred animals under Mexican subtropical conditions is
unknown. However, several phenotypic studies comparing Bos taurus and Bos indicus
crosses under tropical Mexican conditions concluded that F1 heifers calved at younger
age than purebred Bos indicus and F1 animals had higher weaning rates than Bos indicus
(Rios-Utrera et al., 1996; Vega-Murillo et al., 1996). These studies dealt with phenotypic
mean values of groups of animals.
Genetic evaluations of purebred and crossbred animals for economically important
traits (e.g., growth) are needed to genetically improve the cattle population in the
Tamaulipas region. Only recently genetic evaluations in Angus, Charolais and
Simmental breeds have been conducted in Mexico, but included only purebred animals
(Martinez et al., 2002; Rosales-Alday, et al., 2002b; Vega et al., 2002). Conversely,
Mexican commercial producers practice both straight breeding and crossbreeding. Thus,
it becomes necessary to obtain estimates of genetic parameters for these types of herds
under their environmental conditions. Unfortunately, because individual commercial
producers make decisions independently of one another when choosing germplasm, herds
are disconnected and genetic parameters cannot be estimated for the entire population.
The only current alternative is to estimate genetic parameters within herds and to carry
out intra-herd genetic evaluations.
To obtain genetic parameters for growth traits, three independent beef herds under
Mexican subtropical conditions with available growth information agreed to participate in
this study. Herd 1 included purebred Simmental and Simbrah sires mated to S, 3/4S 1/4B,
Simbrah, 1/S 1/ B, 5/16S 11/16B, 1/S %/B and Brahman cows. Herd 2 contained Simmental
and Brahman sires mated to 5/16S 11/16B and Brahman cows. Herd 3 used Simmental and
Brahman sires to mate Brahman cows. Thus, the objectives of this study were to estimate
genetic parameters for birth weight (BWT) and 205 d-adjusted weaning weight (WW205)
for three disconnected herds under Mexican subtropical conditions.
Material and Methods
Data and Herd Management
Data were field records from Simmental, Brahman and crossbred between
Simmental and Brahman calves from three commercial herds in the NE region of Mexico
(Tamaulipas). This region has dry subtropical conditions, with a yearly average
temperature of 23i4C and 812 mm of rainfall. Animals were kept on pastures composed
of star grass (Cynodon nlemfuensis) and native grasses throughout the year. Vitamins
and mineral supplements were provided throughout the year. Heifers were bred to calve
at approximately three years of age. Herds 1 and 2 used natural service during the year.
Herd 3 used artificial insemination with Simmental semen and if cows were not pregnant
after three inseminations, they were mated to Brahman bulls. Cows without pregnancies
in two consecutive years were culled. Calving occurred throughout the year for the three
herds. Calving seasons were classified according to the climatic conditions as follows:
Winter (January to March), Spring (April to June), Summer (July to September) and Fall
(October to December). Records for birth weight and weaning weight, were obtained for
all three herds.
Herd 1 had 2,134 BWT and 1,729 WW205 records from 2,134 calves born in five
years (1994 to 1998). These calves were produced by the mating of 32 Simmental and 32
Simbrah sires to 555 Brahman, 22 1/S %/B, 43 5/16S 11/16B, 288 1/S 1/B, 44 Simbrah, 122
%/S 1/B and 47 Simmental dams. Herd 2 had 1,270 BWT and 373 WW205 records from
1,270 calves born in two years (1995 to 1996). These calves were produced by the
mating of 3 Simmental and 53 Brahman sires to 980 Brahman and 79 5/16S 11/16B cows.
This herd had four calf groups: purebred Brahman and three crosses between Simmental
and Brahman breeds. Herd 3 had 582 BWT and 443 WW205 records from 582 calves
born in four years (1993 to 1996). These calves were produced by the mating of 11
Simmental and 11 Brahman sires to 386 Brahman cows. Number of sires, dams and
calves by sire-breed-group x dam-breed-group combination by herd are shown in Table
Table 3-1. Number of sires, dams and calves by herd and breed-group-of-sire 0I breed-
Breed group of sire
Breed Simmental Simbrah (5/sS 3/8B) Brahman
tBreed group of
tBreed group of
tBreed group or
Herd 1 Simmental 9a
3/4S 1/4B 10
5/16S 11/16B 2
1/4S 3/4B 8
5/16S 11/16B 3
Brahman 11 11
376 1/2S 1/2B 207 B
a Number of sires within breed group of sire x breed group of dam interaction.
b Number of dams within breed group of sire x breed group of dam interaction.
" Number of calves within breed group of sire x breed group of dam interaction.
Calf records without birth date and of unknown parents (sire and dam) were
deleted. Because 205 days was the standard age at weaning in Tamaulipas, weaning
weights were adjusted to 205 days using the Beef Improvement Federation adjustment
procedure (BIF, 1996). Age of dam (AOD) in days was divided into eight year
categories (BIF, 1996). Contemporary groups (CG) were defined as groups of calves
born in the same year, season and had the same sex. Calves of all breed group
combinations were allowed in a CG (Elzo, 1983; Elzo and Famula, 1985; Klei et al.,
1998). At least two sires were required per CG. Because no connection across herds was
present, analyses were made within herds. All pedigree information available was
included in the analyses to reduce bias due to selection and to increase accuracy of
estimation through ties among related animals (Henderson, 1974).
Estimation of Variance Components and Heritabilities
Estimates of variance components were obtained by maximizing the REML log-
likelihood function using the Average Information algorithm (Gilmour et al., 1995),
implemented in the ASREML program (Gilmour et al., 2000). Convergence was assumed
to have been achieved when the log-likelihood changed less than 0.002 in two
A single trait sire-dam model was used to estimate variances. Traits were assumed
to have direct and maternal genetic effects. Because herds had different breed
composition, environmental fixed effects and genetic group effects in the model varied
across herds. Random effects were the same for all traits in the three herds.
Model for Herd 1. Fixed effects for BWT and WW205 were contemporary group
and calf-sex x age-of-dam x breed-group-of-dam interaction. Age of dam was treated as
a continuous variable and expressed in years. Breed group of dam was modeled as a
regression on the Simmental fraction of the dam and were included depending on the
presence or not of the genetic group effect in the crossbred sire, dam or calf. The fixed
regression genetic group effects were: 1) sire intrabreed additive direct, (as a function of
the expected fraction of Simmental alleles in all sires, (pSSIRE)), 2) dam intrabreed
additive direct plus maternal (as a function of the expected fraction of Simmental alleles
in all dams (pSDAhi)), 3) sire interbreed nonadditive direct, as a function of the probability
of Simmental and Brahman alleles at one locus of the progeny of all sires (pS/Bcalf) and
4) dam interbreed nonadditive maternal as a function of the probability of Simmental and
Brahman alleles at one locus of the dam (pS/BDAhf) (Appendix A) (Elzo and Wakeman,
1998). Random genetic effects were sire additive direct, dam additive direct plus
maternal and residual (Table 3-2). In this model, sire random effects contained '/ of the
additive direct genetic effects, whereas dam maternal effect included '/ of the additive
direct genetic effects plus all maternal genetic effects.
The mixed model was:
y = Xec + Xgg + Zss + Zdd + e
y = vector of observations (BWT or WW205).
c = vector of contemporary groups and calf-sex x age-of-dam x breed-group-of-
g = vector of regression group genetic effects (sire intrabreed additive direct, dam
intrabreed additive direct plus maternal, sire interbreed nonadditive direct and dam
interbreed nonadditive maternal),
s = vector random additive direct sire genetic effects,
d = vector of random additive direct and maternal dam genetic effects,
e = vector of residuals,
X, = matrix of 1s and Os that relates calf records to elements of c,
X, = matrix of expected fraction of Simmental alleles that relates calf records to
elements of g,
Zs = matrix of 1s and Os that relates calf records to elements of s and
Zd = matrix of 1s and Os that relates calf records to elements of d.
It was assumed that
y 'Xb ZGZ' + R Z,o,2* A,, Zd ej Add eo2
s2 0 rr2* AZ 2$ A, 0 0
d 0 a Add Zd 0 aj* Add 0
e Jla2 0 0 lae2I
b = [c g]',
X =[X, X,],
Z = [Z Zd ,
C 0,2 A,,* oAddO
0,2, is the sire direct genetic variance,
) is th darlm di~rect plus maternal additive,, genetic varince,
denotes direct matrix product.
The covariance between sire additive direct and dam direct plus maternal genetic
effects was assumed to be zero.
Table 3-2. Fixed environmental effects, fixed genetic group effects, random genetic
effects presented in the model, by herd and trait
Herd 1 Herd 2 Herd 3
Effect BW WW205 BWT WW205 BWT WW205
CG * *
SEX x AODx BGDb *
SEX x AOD *
IAD * *
SIRE * *
DAM * *
aCG = Contemporary group (year x season x sex);
bSEX x AOD x BGD = calf sex x age of dam x breed group of dam interaction.
Model for Herd 2. The model for BWT included the fixed environmental effects
of contemporary group and sex x age-of-dam x breed-group-of-dam interaction. Fixed
genetic group effects were sire intrabreed additive direct, dam intrabreed additive direct
plus maternal and sire interbreed nonadditive direct. Random genetic effects were sire
additive direct, dam additive direct plus maternal and residual (Table 3-2). The model for
WWT included the fixed environmental effects of contemporary group and sex x age-of-
dam interaction. Breed-group-of-dam interaction was not included because this herd had
no calf records from crossbred dams. The only fixed genetic group effect was sire
intrabreed additive direct. Fixed genetic group of dam was omitted after preliminary runs
failed because of confounding (93% of dams in this herd were Brahman). Random
genetic effects were sire additive direct, dam additive direct plus maternal and residual.
age of dam x breed-group-of-dam interaction; SEX x AOD = calf sex x age of dam
interaction; IAD = sire intrabreed additive direct; IAM = dam intrabreed additive direct
plus maternal; INTERBNAD = sire interbreed nonadditive direct; INTERBNAM = dam
interbreed nonadditive maternal (Appendix A).
Model for Herd 3. The model for BWT included the fixed environmental effects
of contemporary group and sex x age-of-dam interaction. Fixed dam genetic group effect
was excluded because all dams were purebred Brahman. Random genetic effects were
sire additive direct, dam additive direct plus maternal and residual (Table 3-2).
Heritabilities for additive direct genetic effects were computed as four times the
sire additive direct genetic variance divided by the phenotypic variance. Heritabilities for
maternal genetic effects were computed as the difference between dam direct plus
maternal genetic variance and the sire additive direct genetic variance divided by the
phenotypic variance. Thus, maternal heritabilities will contain any nonzero covariance
between additive direct and additive maternal genetic effects.
Results and Discussion
Group Genetic Effects
Herd 1. Estimates of regression genetic groups effects for BWT were 1.34 & 0.93
kg for sire intrabreed additive direct, -2.65 A 1.02 kg for dam intrabreed additive direct
plus maternal, -1.59 & 0.88 kg for sire interbreed nonadditive direct and 1.19 & 0.63 kg
for dam interbreed nonadditive maternal. The positive sire intrabreed additive direct
indicates that progeny of Simmental sires had larger BWT than progeny of Brahman
sires, whereas the negative dam intrabreed additive direct plus maternal shows the
opposite, i.e., that dams with more Simmental genes proportion produced smaller calves
at birth than Brahman dams (Table 3-3). Thus, straight bred Simmental dams were
apparently less adapted than Brahman dams to the local environmental conditions of herd
1. Negative sire interbreed nonadditive direct effects indicates that crossbred Simmental-
Brahman calves were smaller than the average of straight bred Simmental and Brahman
calves. Thus, it appears that the local environment prevented the full manifestation of
heterosis for BWT. Contrarily, positive dam interbreed nonadditive maternal effects
suggests that calves of Simmental-Brahman crossbred dams were larger than the average
of straight bred Simmental and Brahman cows (Table 3-3). Thus, Simmental-Brahman
crossbred dams were able to withstand local environmental conditions and allow a fuller
expression of intrauterine growth of crossbred calves than either purebred Simmental or
Table 3-3. Least square means ( A S.E.) for birth weight and weaning weight by breed
group of dam in herd 1
Breed group of n Birth Weight n Weaning Weight
Dam (Kg) (Kg)
Brahman (B) 427 34.60 + 0.17 349 222.7 & 1.7
8/32S 24/32B 40 34.88 & 0.55 31 223.7 & 5.8
10/32S 22/32B 46 34.59 & 0.52 29 193.0 & 6.0
16/32S 16/32B 521 34.32 & 0.16 424 203.6 & 1.6
20/32S 12/32B 81 34.47 & 0.39 74 208.8 & 3.8
24/32S 8/32B 241 33.93 & 0.23 205 211.3 & 2.3
Simmental (S) 41 34.08 & 0.55 38 180.2 & 5.3
For WW205 the estimates of regression genetic group effects for BWT were -28.96
& 8.8 kg for sire intrabreed additive direct, -12.84 & 8.8 kg for dam intrabreed additive
direct plus maternal, 36.18 & 7.7 kg for sire interbreed nonadditive direct and 5.77 & 5.27
kg for dam interbreed nonadditive maternal. Negative sire intrabreed additive direct and
dam intrabreed additive direct plus maternal suggest that Brahman dams were
substantially better mothers than Simmental mothers, probably related to low milk
production by Simmental cows under tropical conditions. Positive sire and dam
nonadditive genetic effects suggest that crossbred calves grew faster and heavier than
straightbred calves, particularly progeny from crossbred dams (Table 3-3). Thus,
adaptability of Simmental-Brahman crossbred calves and dams seemed to be appropriate
to the local climatic, management and nutritional conditions of herd 1.
Herd 2. In this herd, the three genetic group's effects were present for the BWT,
sire intrabreed additive direct, dam intrabreed additive direct plus maternal, sire
interbreed nonadditive direct and just only sire intrabreed additive direct for WW205.
Estimates of regression genetic group's effects for BWT were -0.25 A 1.72 kg for sire
intrabreed additive direct, -0.52 & 1.64 kg for dam intrabreed additive direct plus
maternal and 2.34 & 1.65 kg for sire interbreed nonadditive direct. Negative values for
sire intrabreed additive direct and dam intrabreed additive direct plus maternal indicates
that Simmental sires used in this herd produced smaller calves that Brahman sires and the
Simmental genes in the dam induce to produce smaller calves, this could be due to the
small number of Simmental sires and Simmental-Brahman crossbred dams. Conversely,
positive values for sire interbreed nonadditive direct indicate a better performance for
crossbred calves. For WW205 sire intrabreed additive direct regression coefficient was
80.74 & 76.3 kg, indicating that Simmental sires produce heavier calves at weaning.
Herd 3. In this herd, due to the no presence of crossbred cows, only sire intrabreed
additive direct regression was estimated. For BWT was 1 & 0.27 kg and for WWT was
19.11 & 5.22 kg that suggest a superiority of Simmental sires over Brahman sires, due to
use of artificial insemination in this herd, even though the principal criteria to buy semen
was the price, not the genetic value of the sire.
Variance Components and Heritabilities
Herd 1. Estimates of variance components and heritabilities for BWT and WW205
are shown in Table 3-4. The genetic variance for BWTD was half as large as that for
BWTM, indicating that maternal genetic effects were more important than direct genetic
effects in Herd 1. Heritability estimates for both traits were low probably in part due to
the practice of utilizing imported Simmental and Simbrah sires of low predicted genetic
values for BWT and partly due to the impact of the tough environmental conditions on
animals in herd 1. Thermal stress, ticks and low forage quality appear to have prevented
both purebred and crossbred animals from expressing their genetic potential for BWT.
The small genetic variability expressed for BWT direct and maternal suggests that, under
the current environmental conditions of herd 1, response to selection for BWT will be
Genetic variances and heritabilities for WW205D direct were larger than for
WW205M. Calves' direct ability to grow was more important than the maternal
environment between birth and weaning. Thus, selection for WW205D is likely to yield
larger responses than selection for WW205M in this herd.
The heritability for WW205D was within the range reported by Mohiuddin (1993)
for Simmental in the USA, the UK and Australia. It was close to the unweighted and
weighted values given by Koots et al. (1994). The heritability of WW205M in herd 1
was in the upper limit of the values reported by Mohiuddin (1993) for Simmental in the
USA, the UK and Australia. It was larger than estimates for Brahman in the USA and
Australia (Mohiuddin, 1993).
Herd 2. Estimates of variance components and genetic parameters for BWT and
WW205 in herd 2 are shown in Table 3-4. The extreme unbalancedness of this dataset
may have caused problems of confounding and multicollinearity. Therefore estimates of
variance components and genetic parameters from this herd should be viewed with
extreme caution. The sire additive direct genetic variance was larger than the dam
additive direct plus maternal genetic variance. This resulted in a negative variance for
additive maternal genetic effects. The extreme unbalancedness of the dataset from herd 2
(Table 3-1) may have contributed to these poor estimates of additive genetic variances for
BWT. The additive maternal genetic variance and the maternal heritability were set to
zero (Table 3-4).
Table 3-4. Estimates of genetic variances and heritability ( A S.E.) for birth weight
(BWT) and weaning weight (WW205) in all herds
Traita Genetic variances Heritability
BWTD 0.70 + 0.45
BWTM 1.33 & 0.05
BWTP 14.63 & 0.46
WW205D 191.50 & 69.80
WW205M 96.36 & 9.46
WW205P 847.30 & 4.53
BWTD 6.00 & 1.53
BWTP 7.16 &0.44
WW205D 525.40 & 229.00
WW205M 270.20 & 179.90
WW205P 825.90 & 41.74
BWTD 0.16 &0.42
BWTM 0.93 & 0.44
BWTP 6.19 &0.38
WW205D 250.10 & 180
WW205M 206.70 & 119.90
WW205P 1161.83 & 86.79
" D = direct, M = maternal, P = phenotypic
b Negative maternal variance set to zero. Heritability set to zero.
0.05 & 0.03
0.09 & 0.03
0.23 & 0.08
0.11 & 0.04
0.84 & 0.17
0.63 & 0.28
0.33 & 0.22
0.03 & 0.07
0.15 & 0.07
0.22 & 0.15
0. 17 & 0. 10
Genetic variances and heritabilities for WWT traits were very large (but within the
parameter space), indicating again a possible problem of confounding in this dataset. The
heritability for WW205D was twice as large as the heritability for WW205M, which
indicates that additive direct genetic effects were more important that maternal genetic
effects in herd 2. This result is in agreement with the results for these traits in herd 1.
The heritabilities for WW205D traits were not within the range reported by Mohiuddin
(1993) for Simmental in the USA and the UK. Conversely, the heritability for BWTD
was similar to the value reported by the same author for Brahman in Australia.
Herd 3. Table 3-4 shows the estimates of variance components and genetic
parameters for BWT and WW205 for herd 3. The heritability for BWTD was very small
(0.03) and five times smaller than the heritability for BWTM (0.15). This suggests a
strong maternal effect over birth weight that could be due to all dams in herd 3 were
Brahman. Again, the extremely low BWTD heritability may have been partly due to
using semen from sires with low predicted genetic values for this trait (as in herd 1) and
partly because of difficult environmental conditions.
As with herds 1 and 2, the heritability for WW205D was much greater (almost
twice) than the heritability for WW205M, again indicating that additive direct genetic
effects were more important than additive maternal genetic effects for WWT under the
subtropical conditions of the Tamaulipas region. Similarly, these heritabilities for WWT
traits suggest a positive response to selection for weaning weight in this herd, both direct
Heritabilities for BWT in herds 1 and 3 were substantially smaller than those
obtained in the Romosinuano-Zebu (Turipana, Colombia, Elzo et al., 1998a),
Sanmartinero-Zebu (La Libertad, Colombia, Elzo et al., 2001) and Angus-Brahman
(Florida, USA, Elzo and Wakeman, 1998) multibreed herds under subtropical conditions
both for direct (. 16 to .30) and for maternal genetic effects (. 14 to .29). This suggests that
environmental conditions of these three herds were less severe than those encountered in
herds 1 and 3 in Mexico. Therefore animals managed to express more fully their genetic
abilities. Heritabilities for WWTD and WWTM in herds 1 and 3 were larger than those
estimated for Romosinuano-Zebu (.09 to .13) and Sanmartinero-Zebu (.08 to .10) in
Colombia and comparable to those for Angus-Brahman (. 16 to .25) in the USA. This
may be a consequence of the practice of choosing imported sires with low predicted
genetic values for BWT, this limits the genetic variability among these sires for BWT,
but not for WWT, as evidenced in heritability estimates for WWTD in both herds.
Heritabilities for BWT and WWT traits for herd 2 were vastly different from those
obtained in these three Bos taurus -Bos indicus multibreed herds. However, herd 2 was
extremely unbalanced and may have had problems of confounding and multicollinearity
as indicated above.
Genetic parameter estimation using data from commercial herds is not easy because
there are factors that may not be feasible to include in a statistical model, such as owner
criteria to define management conditions. The three herds included here were subjected
to different environmental and management conditions, even though they had the same
breeding objective; to recreate the Simbrah breed. Genetic parameter estimates were
different across herds. Only variances and genetic parameter estimates from herds 1 and
3 were relatively similar. Better estimates of variance components and genetic
parameters could have been obtained had these three herds been connected by using
common sires. This would have permitted estimation of genetic parameters based on
records from all three herds, thus increasing the total number of animals and observations
and reducing standard errors of variances and genetic parameters.
Weaning weight is an economically important trait because the obj ective of
Mexican subtropical cattle production systems is to produce weaning calves for export to
the USA or for sale to regional feeders. Thus, additive direct and maternal genetic effects
for weaning weight are important to measure and to use for sire evaluation in this region.
The results obtained here for the three herds indicate that weaning weight (direct and
maternal) showed enough genetic variation to be used to implement intra-herd selection
programs. Additional genetic progress could be achieved if herds were genetically linked
through use of common sires. It would be advantageous for commercial producers to
consider the use of common sires as part of their genetic management. Other traits, such
as yearling weight, carcass and survival traits, all of which are economically important,
should also be included in future genetic evaluations and improvement programs.
Variance estimates and heritabilities for pre-weaning weight traits in three
commercial herds in Tamaulipas showed that sufficient genetic variation existed to
justify the implementation of within-herd genetic improvement programs, particularly for
weaning weight traits. To have an even better chance of success for genetic
improvement, herds need to be linked through the use of common sires. This would
improve accuracy for both genetic prediction of genetic values and estimation of genetic
parameters. Therefore increased genetic progress for weight traits should be expected. It
would also be important to implement regional and(or) national genetic improvement
programs that can use information from purebred and crossbred animals.
Purebred and crossbred Simmental and Brahman records from three unconnected
herds were used to estimate within-herd variance components for additive direct and
maternal genetic effects for birth weight and 205 d-adjusted weaning weight. Herds were
located in the subtropical region of Tamaulipas, Mexico. Variances were estimated by
restricted maximum likelihood procedures using a sire-dam mixed model. Fixed effects
were contemporary group, age of dam x sex of calf x breed group of dam interactions.
Genetic group effects were sire intrabreed additive direct, dam intrabreed additive
maternal, sire interbreed Simmental-Brahman nonadditive direct and dam interbreed
Simmental-Brahman nonadditive maternal. Random genetic effects were sire additive
genetic direct, dam additive genetic direct plus maternal and residual. Effects in the
model varied according to trait and breed composition of the herd. Within-herd
heritability estimates were 1) 0.05 & 0.03 for BWTD, 0.09 & 0.03 for BWTM, 0.23 & 0.08
for WW205D, 0. 11 & 0.04 for WW205M in herd 1, 2) 0.84 & 0.17 for BWTD, 0.63 &
0.28 for WW205D and 0.33 & 0.22 for WW205M, in herd 2 and 3) 0.03 & 0.07 for
BWTD, 0. 15 & 0.07 for BWTM, 0.22 & 0. 15 for WW205D and 0. 17 & 0. 10 for
WW205M. These intra-herd heritabilities showed more genetic variability was expressed
for weaning weight than for birth weight traits. Thus, more genetic progress could be
achieved for weaning weight direct and maternal than for birth weight traits in these
herds. As more data is collected in these three herds, more precise estimates of genetic
parameters and more precise selection for weight traits will be achieved.
GENETIC PARAMETERS AND TRENDS FOR PREWEANING GROWTH TRAITS
IN THE MEXICAN SIMMENTAL POPULATION
The Simmental breed has spread throughout the world and its total numbers now
approach 41 million animals (BSCSL, 2003). The breed was introduced in Mexico in
1973 with importation of 10 bulls from Germany and Switzerland. In the 1980's large
numbers of animals were imported from the USA and Canada and distributed to all
regions of Mexico; from arid to tropical. Purebred and crossbred Simmental animals are
well accepted by Mexican producers because their offspring are well accepted by both
local and international markets and have good adaptability to a wide range of
environmental conditions. For Mexican cattlemen to consistently produce beef that meet
national and international market demands, it is important to implement national genetic
improvement programs that can identify those animals that most closely meet these
demands. The Office of the Secretary of Agriculture, Livestock and Rural Development
in Mexico recognized the need for a national genetic improvement program in 1998 and
published a document containing steps to be taken to achieve this obj ective (SAGAR,
1998). This document was written with participation of representatives from breed
associations, national research institutes and universities. One of the steps was the
genetic evaluation of purebred and crossbred beef animals using field data. This work
required estimation of genetic parameters for the traits of interest in the populations to be
evaluated. In addition, to evaluate the impact of selection decisions on the population,
genetic trends would be needed. Thus, the obj ectives of this paper were the estimation of
additive direct and maternal genetic variances and covariances, heritabilities, genetic
correlations and genetic trends for birth weight and weaning weight in the Mexican
Material and Methods
Data and Pedigree Files
Birth weights (BW) and weaning weights (WW) from Simmental calves were
provided by 434 herds belonging to the Mexican Simmental-Simbrah Association. Herds
were located in all regions of Mexico from arid to tropical (Fig. 4-1). Data were
collected from 1982 to 1999.
Figure 4-1. Distribution of Simmental herds in Mexico by state.
Data editing was done according to the Beef Improvement Federation
recommendations (BIF, 1996). Records outside three standard deviations were discarded
as outliers. Calves with unknown parents and calves without BW and WW records were
deleted. Four calving seasons were defined: winter (January-March), spring (April-June),
summer (July-September) and fall (October-December). Contemporary groups were
defined as animals born in the same herd, year and season. Sex was not included as part
of the contemporary group definition because of insufficient data. However, sex was
included as part of the mixed model. Contemporary groups with less than 10 calves
and(or) less than two sires were eliminated. Alphanumeric animal identifications were
renumbered with a FORTRAN program (Misztal, 1993). Contemporary group
connectedness through common sires was verified with a FORTRAN program (Elzo,
2002). Ten percent of the original data were not connected and were eliminated. The
connected dataset contained 18,383 BW and 9,023 WW from 18,460 calves. These
calves were the progeny of 2,152 sires and 1 1,334 dams. A pedigree file was constructed
that included calf, sire and dam. Only calves with purebred parents or with more than
31/32 Simmental were included in the pedigree file. The pedigree file included 26,539
animals. A fixed effects analyzes was made for both variables, BW and WW, including
sex, season, age of dam in years and year of calf s birth.
Estimation of Genetic Parameters
Genetic variances and covariances were estimated by Restricted Maximum
Likelihood using a derivative-free algorithm (Graser et al., 1987). Computations were
performed using the MTDFREML package (Boldman et al., 1995). A two-trait (BW,
WW) animal model with additive direct and maternal genetic effects was used to estimate
covariance components and to predict genetic values. Because dams had repeated
records, permanent environment effect of the dam was included. Fixed effects were
contemporary group (BW, WW), sex (BW, WW), age of dam (in days) at calving linear,
quadratic and cubic (BW, WW) and age of calf (in days) at weaning linear, quadratic and
cubic (WW only). Random effects for BW and WW were animal additive direct, dam
additive maternal, dam permanent environment maternal and residual. In matrix
notation, the two-trait mixed linear model can be represented as follows:
y = Xb + Zi ui + Zd Ud + Zpe dpe+ e
E[y] = Xb
u, G ,,, *A G,, ")d*A 0 0
ud G()d A G()dd *A 0 0
de 0 0 Reave *I 0
e 0 0 0 Rose*
y = vector of BW and WW records,
b = vector of contemporary groups (herd-year-season), sexes and age of dam
covariates (linear, quadratic and cubic) for BW and WW and calf age at weaning
covariates for WW only (linear, quadratic and cubic) fixed effects,
ui = vector of random calf additive direct genetic effects ('/ sire additive direct
plus Mendelian sampling from the sire and the dam of the calf),
Ud = vector of random dam additive maternal genetic effects ('/ dam additive direct
dpe = vector of random dam permanent environment maternal effects,
e = vector of residuals,
X = matrix of is, Os and linear, quadratic and cubic factors that relates calf records
to elements of b,
Zi = matrix of 1s and Os that relates calf records to elements of ui,
Zd = matrix of 1s and Os that relates calf records to elements of ud and
Ze = matrix of 1s and Os that relates animal records to elements of dpe.
For the definitions of covariance matrices, = direct product, A = additive
relationship matrix and the sub scripts i = calf, d = dam, 1 = BW and 2 = WW. Thus,
Gonl = :11,11 ~11,12121 = 2 x 2 matrix of covariances between calf additive direct
genetic effects for trait j and calf additive direct genetic effects for trait j' (j, j' = 1, 2),
G~ld 1,d1 1,d2 = 2 x 2 matrix of covariances between calf additive direct
G 0dIld12,d1 12,d2
genetic effects for trait j and dam additive maternal genetic effects for trait j' (j, j' = 1, 2),
G d1di d1,11 d1,12d21 = 2 x 2 matrix of covariances between dam additive maternal
genetic effects for trait j and calf additive direct genetic effects for trait j' (j, j' = 1, 2),
Gddr d1,d1dd d1~dd2,d = 2 x 2 matrix of covariances between dam additive maternal
genetic effects for trait j and dam additive maternal genetic effects for trait j' (j, j' = 1, 2),
R~epel,pelO 0ee = 2 x 2 matrix of covariances between dam permanent
environment maternal effects for trait j and dam permanent environment maternal effects
for trait j' (j, j' = 1, 2),
Ro =,,, = 2 x 2 matrix of covariances between temporary
environmental effects for trait j and temporary environmental effects for trait j' (j, j' = 1,
Calf additive direct and dam additive maternal genetic values were predicted using
the same direct-maternal animal model and MTDFREML software (Boldman et al.,
1995) used to estimate covariance components and genetic parameters. Yearly means of
predicted calf and dam genetic values were subsequently computed for BW and WW.
Calf and dam genetic trends were generated by plotting BW and WW over time (Smith,
Results and Discussion
All effects included in the model were statistically significant (P<0.01). Model
explained 4% of the variance for birth weight and 5% for weaning weight. Means for
birth weight and weaning weight by calf sex, season and age of dam are presented in the
Table 4-1. Birth weight and weaning weight general means were 38.7 & 5.6 kg and 244.9
& 45.8 kg, respectively. Birth weight mean is similar to that reported by Cundiff et al.
(1998), where several breeds were evaluated including Gelbvieh, which have a similar
performance as Simmental. In this study BWT mean is similar with that reported for the
Simmental breed by the BIF (1996). For Mexican Simmental females in this study, there
is a superiority of 2.4 kg with respect to the BIF (1996) report. Means for calving season
do not show a clear difference because the data came from different Mexican regions,
including tropical and subtropical conditions where seasonality is not present. The higher
mean value for birth weight was for ten years old dams and the smallest was for three
years old dams. There is a discrepancy with the BIF (1996) report, where the mature age
is between 5 and 9 years.
Table 4-1. Least square means ( + S.E.) for birth weight and weaning weight by calf sex,
season and age of dam for Mexican Simmental breed
Effect N Birth Weight n WW205-d
Total 19,367 38.82 f 0.04 8,800 235.22 f 0.50
Males 9,720 40.34 f 0.13 5,460 244.31 f 1.22
Females 9,647 38.68 f 0.12 3,340 235.63 f 1.19
Jan-Mar 5,802 39.86 f 0.13 2,298 243.30 f 1.41
Apr-Jun 5,645 39.82 f 0.13 2,708 245.54 f 1.33
Jul-Sep 4,322 39.39 f 0.14 2,091 234.48 f 1.39
Oct-Dec 3,598 39.95 f 0.14 1,703 236.55 f 1.46
Age of dam
(years) 2 2,592 39.44 f 0.12 1,087 241.01 f 1.56
3 2,684 38.69 f 0.15 1,270 235.64 f 1.51
4 2,960 39.20 f 0.14 1,343 238.24 f 1.48
5 2,578 39.43 f 0.15 1,264 235.75 f 1.53
6 2,318 39.68 f 0.15 1,206 240.85 f 1.56
7 1,927 39.62 f 0.16 1,006 238.99 f 1.66
8 1,371 39.74 f 0.18 701 243.17 f 1.96
9 1,063 39.81 f 0.20 529 240.47 f 2.20
10 369 39.85 f 0.30 162 243.68 f 3.80
11 240 39.70 f 0.37 109 235.19 f 4.61
12 265 39.40 f 0.35 123 246.66 f 4.35
The weaning weight mean of 23 5.20 kg is higher than the range of 167-182 kg, for
several crossbred groups under Mexican tropical conditions (Rios-Utrera et al., 1996).
Unfortunately they did not use the Simmental breed. Comparing with the USA
Simmental breed, the mean found in this study was smaller than the 265.9 kg mean
reported by Lee et al. (1997) and is similar to the reported by Szab6 et al. (2002) in
Phenotypic trend for birth weight is -129 grams per year and is highly significant (P
< 0.001) (Figure 4-2). It looks like Mexican producers intuitively have been selected
sires and dams to reduce the birth weight in their herds, to avoid calving problems.
Phenotypic trend for weaning weight is -2 kg per year and is highly significant (P <
0.001) (Figure 4-3).
1980 1982 1984 1986 1988 1990 1992 1994 1996
Figure 4-2. Birth weight by birth year of Mexican Simmental breed.
1990 1992 1994 1996 1998
Figure 4-3. Weaning weight by birth year of Mexican Simmental breed.
Estimation of Genetic Parameters
Parameters estimates of covariance components and genetic parameters for BW and
WW direct and maternal are shown in Table 4-2. The additive genetic variance for BW
direct was three times larger than for BW maternal, yielding heritability estimates of 0.40
for BW direct and 0.12 for BW maternal. Similarly, the additive genetic variance for
WW direct was twice the size of WW maternal, producing estimates of heritability of
0.33 for WW direct and 0.19 for WW maternal. These heritabilities were almost
identical to those used in the 2003 Simmental Sire Summary for the US Simmental
population (0.39 for BW direct, 0.28 for WW direct, 0.16 for WW maternal; ASA, 2003).
Phenotypic variances were 19.94 kg2 for BW, 690.17 kg2 for WW. The similarity of the
Mexican and US heritability estimates suggests that Mexican producers imported a
representative sample of the American Simmental population during the period covered
in this study (1982 to 1999).
Table 4-2. Covariance components and genetic parameters for birth weight and weaning
weight in the Mexican Simmental population
Traitb BWD BWM WWD WWM
BWD 0.40 (7.91) -0.63 0.73 0.00
BWM -2.72 0.12 (2.40) -0.02 0.00
WWD 31.09 -0.45 0.33 (228.63) -0.39
WWM 0.00 0.00 -66.06 0.19 (128.05)
"Heritabilities (genetic variances in kg2 in parenthesis) on the diagonal, genetic correlations above
the diagonal and genetic covariances (in kg ) below the diagonal.
bBWD = birth weight direct: BWM = birth weight maternal: WWD = weaning weight direct:
WWM = weaning weight maternal.
The negative correlations between additive direct and maternal genetic effects
found for BW (-0.63) and WW (-0.39) in the Mexican Simmental population were also
comparable to US Simmental values (ASA, 2003; Elzo et al., 1986; Garrick et al., 1989;
Lee et al., 1997). Across-trait additive genetic correlations were all zero or near zero
except for the correlation between BW direct and WW direct (0.73); again similar to their
corresponding estimates in the US Simmental population. The phenotypic covariance
between BW and WW (30.87 kg2) and the phenotypic correlation (0.26), were both
comparable to values obtained in US Simmental (Elzo et al., 1986; Garrick et al., 1989).
Permanent environmental maternal effects were not important for BW (permanent
environmental variance = 0.81 kg2 and ratio of permanent environmental variance to
phenotypic variance = 0.04). Permanent environmental effects were more important for
WW (permanent environmental variance = 6.86 kg2 and ratio of permanent
environmental variance to phenotypic variance = 0.10). This permanent environment
ratio of 0. 10 for WW was analogous to the estimates in the US for bulls (0. 12) and for
heifers (0.09) as reported by Lee et al. (1997). As in the US Simmental population,
permanent environmental maternal effects in Mexican Simmental population were
substantially less important than either additive direct or maternal genetic effects.
Heritabilities and genetic correlations estimated for preweaning traits in Mexican
Simmental population suggest that selection for additive direct genetic effects for BW
and WW in the appropriate direction (e.g., smaller BW and larger WW) should be
successful. Selection for WW maternal also appears feasible. However, the selection
program should monitor genetic predictions for direct and maternal effects and take steps
to lessen the impact of the negative correlations between additive direct and maternal
genetic effects for BW and WW. One such measure could be to select sires in two steps.
First, define a minimum value of sire expected progeny difference for WW maternal.
Second, within the set of sires that meets this criterion, choose sires whose expected
progeny difference for direct growth is appropriate to the environmental conditions where
their progeny will grow.
Trends for predicted calf additive direct and dam additive maternal genetic values
are shown in Figure 4-3 for BW and in Figure 4-4 for WW. Trends for calf additive
direct genetic values were flat between 1981 and 1988 and tended to increase from 1989
to 1999 for BW (0.25 kg) and WW (1.40 kg). These upward trends for calf BW and WW
direct suggest that, since 1988, Mexican cattlemen paid more attention to expected
progeny differences (probably for WW direct) when purchasing semen or sires from the
US Simmental population. In contrast, the US Simmental population increased their WW
direct by approximately 10 kg while maintaining their BW direct constant (ASA, 2003).
Trends for dam additive maternal genetic values tended to decrease from 1981 to
1999 for both BW (-0.07 kg) and WW (-0.02 kg). These small negative trends may be a
consequence of Mexican cattlemen choosing US sires primarily for their own ability to
grow (i.e., for their additive direct genetic predictions), particularly for WW, regardless
of their prediction for maternal genetic effects. If sires chosen during this period had
negative predicted values for maternal genetic effects or maternal "milk" (ASA, 2003),
negative BW and WW genetic trends for dam maternal genetic effects could occur
because of negative correlations between direct and maternal genetic effects. During the
same period, the US Simmental population increased their dam additive maternal genetic
values by approximately 6 kg (ASA, 2003).
Figure 4-4. Trends for birth weight calf additive direct (BWD) and dam additive maternal
(BWM) predicted genetic values in the Mexican Simmental population.
0 .4 -1 + WW M
1 i l l a i l l i n
a. -0.05 -
1982 1984 1986 1988 1990 1992 1994 1996 1998 2000
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000
Figure 4-5. Trends for weaning weight calf additive direct (WWD) and dam additive
maternal (WWM) predicted genetic values in the Mexican Simmental
The amounts of genetic variability found in the Mexican Simmental population
suggest that there is ample opportunity for genetic improvement for preweaning direct
and maternal growth traits. A national genetic evaluation program followed by judicious
use of genetic predictions and sound-mating strategies should yield substantially higher
genetic trends than those found between 1982 and 1999.
The large amounts of genetic variability estimated for birth weight and weaning
weight direct and maternal traits suggest that the Mexican Simmental population could
quickly respond to selection programs to improve weaning weights. Care should be
taken to choose sires with low genetic predictions for birth weight direct to avoid calving
difficulties. Further, selected sires should have an appropriate predicted genetic value for
weaning maternal genetic effects, otherwise milk yield could decrease substantially
because of the negative genetic correlation between direct and maternal effects. Other
traits such as yearling weight, carcass traits and reproductive traits should also be
considered in future genetic evaluations in order to have a more comprehensive genetic
Birth weight (18,383) and weaning weight (9,023) records from Mexican
Simmental beef cattle population were used to estimate genetic additive direct and
maternal variances, heritabilities, genetic correlations and genetic trends. Information
was collected from 549 herds owned by members of the Mexican Simmental-Simbrah
Association between 1982 to 1999. Contemporary groups were defined as herd-year-
seasons. A two-trait animal model with additive direct and maternal effects was used to
estimate genetic variances and genetic parameters and to predict genetic values.
Heritabilities for birth weight were 0.40 for additive direct and 0.12 for additive maternal
and the additive direct-maternal genetic correlation was -0.63. Weaning weight
heritabilities were 0.33 for additive direct and 0.19 for additive maternal and the additive
direct-maternal correlation was -0.67. The only across-trait nonzero genetic correlation
was between birth weight direct and weaning weight direct (0.73). Maternal heritabilities
were lower than direct heritabilities and negative correlations between additive direct and
maternal effects existed for both traits. Genetic trends in the Mexican Simmental
population were positive for additive direct and negative for additive maternal genetic
effects. These trends suggest that most of the selection emphasis was likely placed on
additive direct genetic effects for weaning weight. Because of the negative correlation
existing between additive direct and maternal genetic effects in the Mexican Simmental
population, selection programs for preweaning growth will need an appropriate balance
between these two types of effects.
MATERNAL LINE EFFECT INT PRODUCTIVE TRAITS INT MEXICAN
Environmental effects for many economically important traits in beef cattle include
the maternal environment provided by the dam. A maternal effect is any influence, other
than the contribution resulting from nuclear genes, the dam has on the phenotype of her
progeny (Rohrer et al., 1994). Since mitochondria play an important role in intracellular
protein and energy metabolism, the assumption that cytoplasmic effects could be
involved in productive traits is reasonable (Bell et al., 1985). Differences among
maternal cytoplasmic lines in dairy cattle account for 2% of the total variation for milk
yield and 3.5% of total variation for fat percentage (Bell et al., 1985). Cytoplasmic effect
was a significant source of variation for open days (Tess et al., 1987). Huizinga et al.
(1986) found that cytoplasmic variation accounted for 5.6, 4.8, 6.2, 10.1 and 2.5 of the
total variation in milk yield, fat percentage, protein percentage, kilograms of fat plus
protein and milk return in dollars, respectively.
Cytoplasmic effects in beef cattle accounted for 2% of the variation in birth weight,
5% in average daily gain and 5% in weaning weight adjusted to 205 days in Hereford
calves (Tess et al., 1987). Brown et al. (1988) found non important correlations between
mitochondrial metabolism and weaning weight and yearling growth in beef cattle.
Cytoplasmic effects were not detected for birth weight, weaning weight and post weaning
average daily gain in Brangus cattle (Rohrer et al., 1994). In sheep no cytoplasmic
effects were found in the Columbia and Targhee breeds for birth weight, weaning weight,
fleece weight and number of lambs born (Hanford et al., 2003, Van Vleck et al., 2003).
Cytoplasmic line could be used as an alternative for selection under tropical
conditions, where Bos indicus x Bos taunts crosses have been used to increase production
traits (Olson et al., 1990; Elzo et al., 2001; Dominguez et al., 2003) those studies indicate
a superiority of the crossbred animal over the purebred animals both Bos taunts or Bos
indicus, for productive traits. Roberson et al. (1986) found differences between
reciprocal crosses for birth traits. A difference of more than 7 kg for birth weight has
been observed between reciprocal cross embryos regardless of breed type of surrogate
dam (Baker et al., 1990; Thallman et al., 1992). Cytoplasmic inheritance could be
responsible for some of these differences.
In Mexico producers are concerned about dam selection and because dams usually
produce less than 10 calves in their lives, the prediction of breeding values for dams is
inaccurate. In some cases enough information for a reliable genetic evaluation of cows
becomes available, but too late in the life of the animal to be useful for selection.
However, by evaluating dams as a part of a cytoplasmic line group, breeders may have
enough information on the lines to detect those that are genetically superior. This
information would be included in genetic strategies to improve economically important
beef traits. Therefore, the objectives of this study were to determine the effects of
cytoplasmic line effect (CY) on birth weight and weaning weight in Mexican Simmental
and Simbrah populations.
Material and Methods
Data and pedigree files
The data used in the study were described previously (Rosales-Alday et al., 2004a).
All animals were traced to their cytoplasmic origin which was defined as the foundation
female in the maternal line of the pedigree. The oldest registered dam in the pedigree
was used as the origin of the cytoplasmic line. The Mexican Simmental-Simbrah
Association started recording data in 1982 and included some animals born in the early
50's. In some instances maternal lines were traced back for seven generations. A
cytoplasmic line was defined as all animals sharing a common cytoplasmic source (Tess
et al., 1987). Traits considered were birth weight (BWT) and weaning weight adjusted to
205 days (WW205) according to BIF recommendations (1996).
Three data sets were constructed. The first included the entire data set including
both purebred and crossbred animals. This set included 36,292 records for BWT and
11,314 for WW205. The second included only crossbred animals (31,607 records for
BWT and 8,936 for WW205), while the third included only purebred Simmental (3,447
records for BWT and 1,492 for WW205). For each data set connectivity for sires
throughout contemporary groups was verified and information without any genetic
connection was deleted. Only lines with ten or more recorded calves were included in the
analysis. A total of 1,680 different maternal lines were used in this in this data set.
Another restriction was the contemporary group. Sire groups with less than ten records
Estimation of Genetic Parameters
Genetic variances and covariances were estimated by Average Information
Restricted Maximum Likelihood (Gilmour et al., 1995, 2002). The computational
package used to calculate (co)variance components was ASREML (Gilmour et al. 2000).
A single trait BWT or WW205, sire-maternal grand sire model (SMGS) with additive
direct, additive maternal and cytoplasmic line effects was used. Fixed effects were sex,
age of dam in days, linear and quadratic, contemporary group and sex x age of dam
interaction. Covariances between direct and residuals, maternal and residual and
cytoplasmic and residual effects were assumed to be zero. Random effects were sire
additive direct, maternal grand sire additive direct, maternal cytoplasmic line effect and
residual. The mixed random model for each trait can be represented as follows:
y = Xb + Zi ui + Zm um + Zcy Ucy +
y = vector of observations (BWT or WW205),
b = vector of sex, age of dam in days, linear and quadratic, contemporary group and
sex x age of dam interaction,
ui = vector of additive direct genetic effects,
um = vector of additive maternal genetic effects,
Ucy = vector of maternal cytoplasmic line effects,
e = vector of residuals,
X = matrix of 1s and Os that relates calf records to fixed effects,
Zi = matrix of 1s and Os that relates calf records to additive direct genetic effects,
Zm = matrix of 1s and Os that relates calf records to additive maternal genetic
Zcy = matrix of 1s and Os that relates calf records to line of cytoplasmic origin
It was assumed that
--XGX' +R Z2a2*A Zm 0*A Zc y2*A lef
uO 2 A Z', e2 A Z m,, Z', *A Z cyc Z', *A 0
us ~ M7VN 0 a AZ'_ Z,"m,, Z'm~ *A o *A Zcy m,~~ Z'"' *A 0
ucy 02 *AZ' Z ",,Z'C *A Z cy Z'c *A "2 *A 0
e 0 le 0 0 0 le)
G2i = is the sire direct genetic variance,
G2m = is the maternal grandsire direct genetic variance,
G2cy = is the cytoplasmic variance,
G2e = TOSidual error variance,
oi,m = covariance between direct and maternal effect,
Gi,cy = COVariance between direct and cytoplasmic effect,
Gm,cy = COVariance between maternal and cytoplasmic effect,
A = relationship matrix,
I = identity matrix.
Results and Discussion
Estimation of Genetic Parameters
Purebred and crossbred data. Estimations of covariance components and genetic
parameters for BWT direct, maternal and cytoplasmic line for the complete data set,
purebred plus crossbred, are shown in Table 5-1. The covariance between BWTD and
cytoplasmic line was deleted from the model for BWT because the standard error
exceeded the parameter value. While the heritability for BWTD approximated that
reported by Rosales-Alday et al. (2004a), heritability for BWTM was different, 0.04 vs.
0.12, from the previous analysis. Heritability for BWTD is bigger than the parameters
reported by Mohiuddin (1993) and the weighted heritability reported by Koots et al.
(1994). This difference can be explained by the inclusion of information from crossbred
animals and the addition of nonadditive covariance which could bias the estimation of
The ratio of cytoplasmic line variance (0.4 & 0.01 kg2) to total phenotypic variance
(22.04 & 0.3 1 kg2) for BWT was small, 0.02 & 0.003. This indicates that only 2% of the
total variance was explained by the cytoplasmic line effect. This ratio is within the range
reported by others (Bell et al., 1985; Huizinga et al., 1986) who indicated that
cytoplasmic effects accounted for 1 to 4% of the total variation in dairy traits. These
present results are in agreement with the reports by Rohrer et al. (1994), Tess et al.
(1987) and Van Vleck et al. (2003) who found no impact of cytoplasmic effects in beef
cattle and sheep. Those differences could be because in this paper cytoplasmic effect was
included in a Sire-Matemnal Grand Sire model. This model can separate the additive
direct genetic effect, the maternal direct effect and the cytoplasmic effect at the same
The genetic correlation between BWTD and BWTM was 0.30 + 0.08 and different
from the reported values of ASA (2003), Elzo et al. (1986), Garrick et al. (1989), Lee et
al. (1997) and Rosales et al. (2004a) who found negative genetic correlation for those
traits. The correlation between BWTM and cytoplasmic line is 0.52 & 0.43. This
correlation is high because it measures the maternal effect, one from the sire's side and
the other through the cytoplasmic effect. Both have their influence through the dam. But
BWTM measures the additive genetic maternal effect and the cytoplasmic effect
measures the non-nuclear effects. No correlation was found reported in the literature for
Table 5-1. Covariance components and genetic parameters and standard error for birth
weight in the Mexican Simmental-Zebu purebred and crossbred population
Trait 2 Heritability
BIRTH WEIGHT DIRECT 8.57 & 0.65 0.39 & 0.03
rlEIGHT MATERNAL 0.79 & 0.01 0.04 & 0.01
TD,BWTM)a 0.39 & 0.01 0.30 + 0.08 b
TM, CYTOPLASMIC LINE)a 0.3 & 0.24 0.52 & 0.43 b
rlEIGHT CYTOPLASMIC LINE 0.4 & 0.01 0.02 & 0.003"
JG WEIGHT DIRECT 190 & 31.4 0.27 & 0.04
JG WEIGHT MATERNAL 34.2 & 6.4 0.05 & 0.009
JG WEIGHT CYTOPLASMIC LINE 14.7 & 4 0.021 & 0.0060
aBWTD = birth weight direct; BWTM = birth weight maternal
"Cytoplasmic variance/Phenotypic variance.
The covariance components, genetic parameters and standard errors for WW205 in
the Mexican Simmental-Zebu crossbred population are presented in the Table 5-1. All
covariances among those traits were not considered because previous analyses found
them not to differ from zero. Heritability for WW205D for the complete Mexican
Simmental-Zebu crossbred population was 0.27 & 0.04. This value is smaller than those
previously reported by Rosales-Alday et al. (2004a). But they were close to those
parameters reported by Mohiuddin (1993), for American and Canadian Simmental breed.
The ratio of cytoplasmic line variance for WW205 (14.7 & 4 kg2) to total
phenotypic variance (712. 1 + 11.8 kg2) again was small, 0.021 & 0.006. This indicates
that little of the genetic variation existing for this trait is explained by cytoplasmic line
effect on 205d adjusted weaning weight.
Purebred animals. Variance estimates and genetic parameters for purebred
animals for BWT and WW205 are shown in table 5-2. Heritability for BWTD was 0.25
& 0.08 and for BWTM was 0.04 & 0.02. These results differ from those reported by
Rosales et al. (2004a). And the heritabilities were smaller than those reported by
Mohiuddin (1993) and Koots et al. (1994). The main difference is that in the present
study Simmental breed was defined as the breed of the maternal grand dam. This causes
a reduction of the genetic variance which was detected by the analysis.
Table 5-2. Covariance components, genetic parameters and standard errors for birth
weight and weaning weight in purebred Mexican Simmental population
Trait 2 Heritability
BIRTH WEIGHT DIRECT 5.28 & 1.71 0.25 & 0.08
BIRTH WEIGHT MATERNAL 0.88 & 0.34 0.04 & 0.02
BIRTH WEIGHT CYTOPLASMIC LINE 0.35 & 0.19 0.02 & 0.009a
WEANING WEIGHT DIRECT 315.8 & 123.8 0.36 & 0. 13
WEANING WEIGHT MATERNAL 15.11 & 20. 12 0.017 & 0.023
WEANING WEIGHT CYTOPLASMIC LINE 24.6 & 17.9 0.03 & 0.02a
aCytoplasmic variance/Phenotypic Variance.
Heritability for WW205D was 0.36 & 0.13 and it was in agreement with the
parameter reported in previous analyses with the same data (Rosales et al., 2004a). The
heritability found in this analysis was smaller than the reported by Koots et al. (1994) for
WW205D. Heritability for WW205M was 0.017 & 0.023. This parameter is not differing
from zero. This result is different from other reports whose reported heritabilities bigger
than zero (, 1993; Koots et al., 1994; Rosales et al., 2004a). The ratio between
cytoplasmic line and the total phenotypic variance for WW205 was 0.03 & 0.02. This
ratio is in the range reported for cytoplasmic effects in the literature (Tess et al., 1987;
Gibson et al., 1997). No genetic correlations among traits were found for the purebred
Simmental analyses. The number of records in this data set was small, only 3,447
records for BWT and 1,492 for WW205 and account for the large standard errors.
Crossbred animals. Covariance estimates and genetic parameters for Simmental-
Zebu crossbred population for BWT and WW205 are shown in table 5-3. Heritability for
BWTD was 0.44 & 0.03. This parameter was bigger compared with those reported by
Mohiuddin (1993) and heritability reported by Koots et al. (1994). Heritability for
BWTM was 0.03 & 0.01 and it was smaller than the reported for other authors
(Mohiuddin, 1993; Koots et al., 1994; Rosales et al., 2004a). The correlation between
BWTD and BWTM was 0.36 & 0.08, the value of this parameter was different for those
reported by other authors (Elzo et al., 1986; Garrick et al., 1989; Lee et al., 1997; ASA,
2003; Rosales et al., 2004a) who found negative genetic correlation for those traits. The
correlation between BWTM and cytoplasmic line is 0.27 & 0.01 this indicates an
important relation between maternal and cytoplasmic effects.
Heritability for WW205D was 0.31 & 0.05, this value was bigger than those found
by Mohiuddin (1993) and Koots et al. (1994). Heritability for WW205M was 0.05 & 0.01
and was smaller than those reported by Mohiuddin (1993) and Koots et al. (1994). The
ratio between cytoplasmic line and the total phenotypic variance for BWT was 0.02 &
.001 and for WW205 was 0.022 & 0.007. The range of percentage of variation explained
by cytoplasmic line for birth weight and for 205 days adjusted weaning weight was from
2% to 2.2% respectively and they are within the range reported by Bell et al. (1985),
Huizinga et al. (1986) and Tess et al. (1987) in beef cattle and milk yield.
Table 5-3. Covariance components, genetic parameters and SE for birth weight and
weaning weight in Mexican Simmental-Zebu crossbred population
Trait z Heritability
BIRTH WEIGHT DIRECT 9.76 & 0.79 0.44 & 0.03
BIRTH WEIGHT MATERNAL 0.74 & 0.01 0.03 & 0.01
Cov(BWTD, BWTM)a 0.48 & 0.12 0.36 & 0.08b
Cov(BWTM, CYTOPLASMIC LINE)a 0.15 & 0. 11 0.27 & 0.21b
BIRTH WEIGHT CYTOPLASMIC LINE 0.41 & 0.06 0.02 & 0.01"
WEANING WEIGHT DIRECT 212.8 & 37.45 0.31 & 0.05
WEANING WEIGHT MATERNAL 34.42 & 7. 11 0.05 & 0.01
WEANING WEIGHT CYTOPLASMIC LINE 14.62 & 4.61 0.022 & 0.0070
aBWTD = birth weight direct; BWTM = birth weight maternal
"Cytoplasmic variance/Phenotypic variance.
The results found in this paper means that breeders could have another trait to
evaluate the productive performance on cows. With cytoplasmic line effects, breeders
can have more control over birth weight and weaning weight. As a result, the producer
can select lines according to their specific needs. If they need to maintain the birth
weight constant, they can choose maternal lines with cytoplasmic line genetic value for
birth weight close to zero. If they want to increase or decrease the birth weight they can
select the corresponding extreme cytoplasmic line. The same procedure can be utilized
with weaning weight.
The data seem to suggest that variation due to maternal cytoplasm may exist but
has only a small effect on birth weight and weaning weight in the Mexican Simmental-
Zebu population. If cytoplasmic line effects had been greater they may have offered an
additional tool to improve economically important traits such as birth weight and
weaning weight. By combining cytoplasmic line data with EPD's, breeders could
possibly obtain more genetic improvement that if they use only the EPD. Further
investigation is needed in order to determine if selection response could be improved by
including maternal cytoplasmic line in beef cattle selection programs. Even relatively
modest contributions to variation imply substantial differences in performance between
mitochondrial lineages. This may be valuable information when selecting donor cows for
in vitro fertilization, embryo transfer or cloning in the future.
Purebred Simmental and crossbred Simmental-Zebu records were used to estimate
variance components for additive direct, additive maternal and cytoplasmic line genetic
effects on birth weight and weaning weight traits. Records were from the Mexican
Simmental-Simbrah Association. Three data sets were formed; the first included both
purebred Simmental and crossbred animals. The second included only purebred animals.
The third included only crossbred animals. Variances were estimated by Restricted
Maximum Likelihood procedures using a Sire-Maternal Grand Sire model. Fixed effects
were sex, age of dam linear and quadratic contemporary groups and calf-sexxage of dam
interaction. Heritabilities for both purebred and crossbred animals were 0.39 & 0.03, 0.04
& 0.01 for BWTD and BWTM, respectively. The genetic correlation between BWTD
and BWTM was 0.30 + 0.08 and for BWTM and cytoplasmic effects was 0.52 & 0.43.
Heritabilities for WW205D and WW205M were 0.27 & 0.04 and 0.05 & 0.009,
respectively. Heritabilities for Simmental purebred data for BWTD and BWTM were
0.25 & 0.08 and 0.04 & 0.02, respectively. Heritabilities for WW205D and WW205M
were 0.36 & 0.13 and 0.017 & 0.023, respectively. Heritabilities for crossbred animals
were 0.44 & 0.03 and 0.03 & 0.01 for BWTD and BWTM, respectively. The genetic
correlation between BWTD and BWTM was 0.36 & 0.08 and for BWTM and
cytoplasmic effects was 0.27 & 0.21. Heritabilities for WW205D and WW205M were
0.31 & 0.05 and 0.05 & 0.01, respectively. Cytoplasmic line variance and total phenotypic
ratio ranged from 2 to 3% for all traits in all data groups. Modest contributions of the
cytoplasmic mitochondrial effect to birth weight and weaning weight variation were
found in this study. This could be help to select maternal lineages to select cows to be
used as donor for in vitro fertilization, embryo transfer or cloning in the future in
Mexican Simmental population.
Three studies were conducted utilizing data from three commercial herds
maintained under subtropical conditions in Mexico and also with data from the Mexican
Simmental-Simbrah Association to determine genetic parameters for birth weight and
weaning weight. The initial study was conducted with purebred and crossbred
Simmental and Brahman records from three unconnected herds. A Sire-Dam model was
used to estimate within-herd variance components for additive direct and maternal
genetic effects for birth weight and 205 d-adjusted weaning weight. Within-herd
heritability estimates were 0.05 & 0.03 for BWTD, 0.09 & 0.03 for BWTM, 0.23 & 0.08
for WW205D and 0. 11 & 0.04 for WW205M in herd 1; 0.84 & 0. 17 for BWTD, 0.63 &
0.28 for WW205D and 0.33 & 0.22 for WW205M in herd 2 and 0.03 & 0.07 for BWTD,
0.15 & 0.07 for BWTM, 0.22 & 0.15 for WW205D and 0.17 & 0.10 for WW205M for
herd 3. These intra-herd heritabilities showed greater genetic variability was expressed
for weaning weight than for birth weight. Thus, more genetic progress could be achieved
for weaning weight, direct and maternal, than for birth weight traits in these herds. As
more data are collected from these three herds, more precise estimates of genetic
parameters and more precise selection for those weight traits can be achieved.
In the second study, birth weight and weaning weight records from the Mexican
Simmental beef cattle population were used to estimate genetic additive direct and
maternal variances, heritabilities, genetic correlations and genetic trends with a two-trait
animal model with additive direct and maternal effects. Heritabilities for birth weight
were 0.40 for additive direct and 0.12 for additive maternal and the additive direct-
maternal genetic correlation was -0.63. Heritabilities for weaning weight were 0.33 for
additive direct and 0.19 for additive maternal. The additive direct-maternal correlation
was -0.67. The only across-trait nonzero genetic correlation was between birth weight
direct and weaning weight direct (0.73). Maternal heritabilities were lower than direct
heritabilities and negative correlations between additive direct and maternal effects
existed for both traits. Genetic trends in the Mexican Simmental population were
positive for additive direct and negative for additive maternal genetic effects. These
trends suggest that most of the selection emphasis was likely placed on additive direct
genetic effects for weaning weight. Because of the negative correlation that exists
between additive direct and maternal genetic effects in the Mexican Simmental
population, selection programs for preweaning growth will need to strike an appropriate
balance between these two types of effects.
In the third study, purebred and crossbred Simmental and Brahman records were
used to estimate variance components for additive direct, additive maternal and
cytoplasmic genetic effects for birth weight and weaning weight. Records were from the
Mexican Simmental-Simbrah Association. Three data sets were formed. The first
included purebred Simmental and crossbred animals. The second included only the
purebred animals. The third included only the crossbred animals. Variances were
estimated by Restricted Maximum Likelihood procedures using a Sire-Maternal Grand
Sire model. Fixed effects were sex, age of dam linear and quadratic contemporary groups
and calf-sexxage of dam interaction. Heritabilities for purebred and crossbred animals
were 0.39 & 0.03, 0.04 & 0.01 for BWTD and BWTM, respectively. The genetic
correlation between BWTD and BWTM was 0.30 + 0.08 and for BWTM and
cytoplasmic effects was 0.52 & 0.43. Heritabilities for WW205D and WW205M were
0.27 & 0.04 and 0.05 & 0.009, respectively. Heritabilities for Simmental purebred data for
BWTD and BWTM were 0.25 & 0.08 and 0.04 & 0.02, respectively. Heritabilities for
WW205D and WW205M were 0.36 & 0.13 and 0.017 & 0.023, respectively.
Heritabilities for crossbred animals were 0.44 & 0.03, 0.03 & 0.01 for BWTD and
BWTM, respectively. The genetic correlation between BWTD and BWTM was 0.36 &
0.08 and for BWTM and cytoplasmic effects was 0.27 & 0.21. Heritabilities for
WW205D and WW205M were 0.31 & 0.05 and 0.05 & 0.01, respectively. Cytoplasmic
line variance and total phenotypic ratio ranged from 2 to 3% for all traits in all data
groups. Modest contributions of cytoplasmic mitochondrial effect to birth weight and
weaning weight variation were found in this study. This could be helpful in selecting
maternal lineages of cows to be used as donors for in vitro fertilization, embryo transfer
or cloning in the future in Mexican Simmental population.
This series of studies are the first conducted in Mexico. Further genetic evaluation
will be made with the participation of other cattle associations, such as the Angus,
Brangus and Charolais. Mexican cattlemen are requesting this type of research to enable
them to supply a better and competitive product to the regional, national and international
markets. They realize that there are more traits to be evaluated and have agreed to
include yearling weight, scrotal circumference and carcass traits among others. Also
international participation is required to evaluate sires and obtain the expected breeding
values in a wide range of environments like the subtropical conditions of Mexico and
temperate conditions of the USA.
GRAPHIC REPRESENTATION OF NON-ADDITIVE GENETIC EFFECT ON
1. Sire Intrabreed
pSsire=Simmental genes proportion on Sire
2. Dam Intrabreed Additive
Direct plus Maternal
pSDAM=Simmental genes proportion on Dam
(pSsmI~ pBsmT+ pSD>AM PBDAM~
3. Sire Interbreed
pSSIRE=Simmental genes proportion on Sire
pBSIRE=Brahman genes proportion on Sire
pSDAM=Simmental genes proportion on Dam
pBDAM=Brahman genes proportion on Dam
(pSSIR~E PBIDAM4. PSIR~EpSD>AM
4. Dam Interbreed
pSSIRE=Simmental genes proportion on Sire
pBDAM=Brahman genes proportion on Dam
pBSIRE=Brahman genes proportion on Sire
pSDAM=Simmental genes proportion on Dam
LIST OF REFERENCES
Alves Santiago, A. 1976. El cebu: Ganado bovine para los paises tropicales. The zebu:
Cattle for tropical countries. UTEHA, Mexico.
American Simmental Association [ASA]. 2003. Multi-breed international cattle
evaluation. Spring 2003 Simmental/Simbrah sire summary. Online. Available:
http://www. simmental. org/asadata/s03 ss.pdf. Accessed August 20, 2003.
Baker, J. F., C. G. Dorn, and G. A. Rohrer. 1990. Donor, fetal and recipient effects on
birth weight and gestation length of calves. J. Anim. Sci. 68(Suppl. 1):22(Abstr.).
Bans, G., L. R. Schaeffer, E. B. Burnside, 1991. Genetic relationships and linear model
comparisons between United States and Canada Ayrshire and Jersey bull
populations. J. Dairy Sci. 74:1060-1068.
Bans, G., A. Sigurdsson. 1996. Application of contemporary methods for the use of
international data in national genetic evaluations. J. Dairy Sci. 79: 1117-1125.
Barkhouse, K. L., L. D. Van Vleck, and L. V. Cundiff. 1995. Mixed model methods to
estimate breed comparisons for growth and maternal traits adjusted to a 1996 base.
In: Proc. Beef Improvement Federation. 27th Research Symposium and Annual
Meeting. Sheridan, WY.
Bell, B. R., B. T. McDaniel, and O. W. Robison. 1985. Effects of cytoplasmic inheritance
on production traits of dairy cattle. J. Dairy Sci. 68:203 8-205 1.
Beef Improvement Federation [BIF]. 1996. Guidelines for uniform beef improvement
programs (7th Ed). Kansas State Univ., Kansas.
Boettcher, P. J., M. T. Kuhn, and A. E. Freeman. 1996. Impacts of cytoplasmic
inheritance on genetic evaluations. J. Dairy Sci. 79:653-675.
Boldman, K, G., L. K. Kriese, L. D. Van Vleck, C. P. Van Tassel, and S. D. Kachman.
1995. A manual for use of MTDFREML. A set of programs to obtain estimates of
variances and covariances. U. S. Department of Agriculture, Agriculture Research
Broman, K. W. 2002. A tour of R/qtl. Department of Biostatistics, John Hopkins
University. Online. Available:
http://www.biostat.j hsph. edu/~kbroman/qtl/rqtltour.pdf. Accessed January 3 0,
Brown, D. R., S. K. DeNise, and R. G. McDaniel. 1988. Mitochondrial respiratory
metabolism and performance of cattle. J. Anim. Sci. 66: 1347-1354.
British Simmental cattle society Ltd. 2003. Simmental: the dual purpose breed for beef
and milk, both for pure and crossbreeding. Online. Available:
http://www.briti shsimmental.co .uk/breed/index.html Accessed August 25, 2003.
Canavesi, F., D. Boichard, V. Ducrocq, N. Gengler, G. de Jong, and Z. Liu. 2002. An
alternative procedure for international evaluations: Production traits European j oint
evaluation (PROTEJE). Proceedings of the 7th World Congress on Genetics
Applied to Livestock Production. CD-ROM communication no 17-03.
Cundiff, L. V., K. E. Gregory, and R. M. Koch. 1998. Germplasm evaluation in beef
cattle-cycle IV: Birth and Weaning traits. J. Anim. Sci. 76:2528-2535.
De los Santos V. S. G., R. Guarneros A., J. Rosales A., J. W. Holloway and R. D. Randel.
2001. Evaluaci6n productive de germoplasma Tuli y Fleckvieh: Comportamiento
de las crias al destete. Mem. XXIX Reuni6n AMPA. Cd. Victoria, Tamaulipas,
De los Santos V. S. G., J. Rosales A., and M. A. Gonzalez P. 2003. Diagn6stico de la
ganaderia de leche y carne en el municipio de Aldama, Tamaulipas. INIFAP-
SAGARPA. p 74.
Dominguez V. J., R. Nufiez D., R. Ramirez V. and A. Ruiz Flores. 2003. Environmental
effects and repeatability for growth traits in Tropicarne cattle. Tec. Pecu. Mex.
Eccleston, J. A. 1978. Variance components and disconnected data. Biometrics 34:479-
Elzo, M. A. 1983. Multibreed sire evaluation within a country and across countries. Ph.
D. dissertation. Univ. of California, Davis.
Elzo, M. A. 1996. Animal breeding notes. Advanced animal breeding course ANS6386.
Spring 1996. Mimeo. University of Florida, Gainesville.
Elzo, M. A. 1999. Multibreed evaluation theory and application. Seventh Genetic
Prediction Workshop, Kansas City, MO, December 3, 1999.
Elzo, M. A. 2002. Multibreed connectedness program CSET (version 11/21/2001).
Animal Breeding Mimeo Series, No. 54, University of Florida, Gainesville. p 11.
Elzo, M. A., and G. E. Bradford. 1985. Multibreed sire evaluation across countries. J.
Anim. Sci. 60:953-963.
Elzo, M. A., and T. R. Famula. 1985. Multibreed sire evaluation procedures within a
country. J. Anim. Sci. 60:942-952.
Elzo, M. A., E. J. Pollak, and R. L. Quaas. 1986. Genetic parameter estimates for weight
and calving ease traits in the Simmental population. Mimeo. Cornell University,
Ithaca. p 29.
Elzo, M. A., C. Manrique, G. Ossa, and O. Acosta. 1998a. Additive and nonadditive
genetic variability for growth traits in the Turipana Romosinuano-Zebu multibreed
herd. J. Anim. Sci. 76:1539-1549.
Elzo, M. A., R. L. West, D. D. Johnson, and D. L. Wakeman. 1998b. Genetic variation
and prediction of additive and nonadditive genetic effects for six carcass traits in an
Angus-Brahman multibreed herd. J. Anim. Sci. 76: 1810-1823.
Elzo, M. A., and D. L. Wakeman. 1998. Covariance components and prediction for
additive and nonadditive preweaning growth genetic effects in an Angus-Brahman
multibreed herd. J. Anim. Sci. 76:1290-1302.
Elzo, M. A., G. Martinez, F. Gonzalez, and H. Huertas. 2001. Additive, nonadditive, and
total genetic variation and genetic predictions for growth traits in the Sanmartinero-
Zebu multibreed herd of la Libertad. J. CORPOICA 3(2):51-64.
Emmerlin, R., E. A. Mantysaari and M. Lidauer. 2002. Reduced rank covariance
functions for multi-lactation test-day model. Proceedings of the 7th World
Congress on Genetics Applied to Livestock Production. CD-ROM communication
Falconer, D. S. 1989. Introduction to quantitative genetics. 3rd ed. Longman Press,
FAO's Information System on Water and Agriculture. 2003. On line Available:
http ://www.fao.org/ag/agl/aglw/aquastat/dbase/ndxstm. Accessed February 29,
Fries, A.F. 1998. From deviation records to animal models. Proc. 6th World Cong. Genet.
Appl. Livest. Prod., Armidale, Australia, 25:617-620.
Galina, C. S., and G. H. Arthur. 1989. Review of cattle reproduction in the tropics. Part 1.
Puberty and age at first calving. ABA 57:583-590.
Galina, C. S., and G. H. Arthur. 1991. Review of cattle reproduction in the tropics. Part 6.
The male. ABA 59:403-412.
Galina, C. S., and J. M. Russell. 1987. Research and publishing trends in cattle
reproduction in the tropics: Part 1. A global analysis. ABA 55:743-749.
Garrick, D. J., E. J. Pollak, R. L. Quaas, and L. D. Van Vleck. 1989. Variance
heterogeneity in direct and maternal weight traits by sex and percent purebred for
Simmental-sired calves. J. Anim. Sci. 67:2515-2528.
Gengler, N., A. Tijani, and G. R. Wiggans. 1999. Iterative solution of random regression
models by sequential estimation of regression and effects on regression. Online.
Available: http://www. csc.fl/ttn/ccb99/articles/NGengler.pdf. Accessed February
Gibson, J. P., A. E. Freeman, and P. J. Boettcher. 1997. Cytoplasmic and mitochondrial
inheritance of economic traits in cattle. Livest. Prod. Sci. 47: 115-124.
Gilmour, A. R., B. R. Cullis, S. J. Welham, and R. Thompson. 2000. ASREML
Reference manual. Printed by NSW Agriculture, Orange Agricultural Institute,
Forest Road, Orange, NSW, 2800 Australia. p 217.
Gilmour, A. R., R. Thompson, and Cullis B. R. 1995. Average information REML, an
efficient algorithm for variance parameter estimation in linear mixed models.
Gilmour, A. R., R. Thompson, Cullis B. R., and S. J. Welham. 2002. ASREML estimates
variance matrices from multivariate data using the animal model. Proceedings of
the 7th World Congress on Genetics Applied to Livestock Production. CD-ROM
communication no 28-05.
Gosey, J. 1997. Bull genetics: purebreds, composites, fullsibs and half-sibs. Proc. The
range beef cow symposium XV. Rapid City, South Dakota.
Graser, H. U., S. P. Smith, and B. Tier. 1987. A derivative-free approach for estimating
variance components in animal models by restricted maximum likelihood. J. Anim.
Hammond, A. C., T. A. Olson, C. C. Jr. Chase, E. J. Bowers, E. D. Randel, C. N.
Murphy, D. W. Vogt, and A. Tewolde. 1996. Heat tolerance in two tropically
adapted Bos taurus breeds, Senepol and Romosinuano, compared with Brahman,
Angus and Hereford cattle in Florida. J. Anim. Sci. 74:295-303.
Hanford, K. J., G. D. Snowder, and L. D. Van Vleck. 2003. Models with nuclear,
cytoplasmic, and environmental effects for production traits of Columbia sheep. J.
Anim. Sci. 81:1926-1932.
Hanocq, E., Boichard, D., Foulley, J. L. 1996. A simulation study of the effect of
connectedness on genetic trend. Gen. Sel. Evol. 28:67-82.
Henderson, C. R. 1953. Estimation of variance and covariance components. Biometrics.
Henderson, C. R. 1974. General flexibility of linear model techniques for sire evaluation.
J. Dairy Sci. 57:963-972.
Henderson, C. R. 1975a. A rapid method for computing the inverse of a relationship
matrix. J. Dairy Sci. 58:1727-1730.
Henderson, C. R. 1975b. Best linear unbiased estimation and prediction under a selection
model. Biometrics. 31:423-447.
Henderson, C. R. 1976a. Use of relationships among sires to increase accuracy of sire
evaluation. J. Dairy Sci. 58:173 1-173 8.
Henderson, C.R. 1976b. A simple method for computing the inverse of a numerator
relationship matrix used in prediction of breeding values. Biometrics 32:69-83.
Henderson, C. R. 1976c. Inverse of a matrix of relationships due to sires and maternal
grandsires in an inbred population. J. Dairy Sci. 59:1585-1588.
Henderson, C. R. 1976d. Multiple trait sire evaluation using the relationship matrix. J.
Dairy Sci. 59:769-774.
Henderson, C. R., O. Kempthomne, S. R. Searle, and C. M. von Krosigk. 1959. The
estimation of environmental and genetic trends from records subj ect to culling.
Hoffman, M. 1991. How parents make their mark on genes. Science 252: 1250-1251.
Huizinga, H. A., S. Korver., B. T. McDaniel, and R. D. Politiek. 1986. Maternal effects
due to cytoplasmic inheritance in dairy cattle: influence on milk production and
reproductive traits. Livest. Prod. Sci. 15:11-26.
Hutchison, C. A., J. E. Newbold, S. S. Potter, and M. H. Edgell, 1974. Maternal
inheritance of mammalian mitochondrial DNA. Nature, 25 1:53 6-53 8.
Kennedy, K. W. and D. Trus. 1993. Considerations on genetic connectedness between
management units under an animal model. J. Anim. Sci. 71:2341-2352.
Klei, L., R. L. Quaas, E. J. Pollak, and B. E. Cunningham. 1998. Multiple-breed
evaluation. Cornell Animal Science/Extension Document Database [on line]
Available: http://www. simmgene.com/multibre.pdf. Accessed March, 2001.
Koonawootrittriron, S., M. A. Elzo, S. Tumwasorn, and W. Sintala. 2001. Lactation
curves and prediction of daily and accumulated milk yields in a multibreed dairy
herd in Thailand using all daily records. Thai Journal of Agricultural Science
Koonawootrittriron, S., M. A. Elzo, and S. Tumwasomn. 2002a. Multibreed genetic
parameters and predicted genetic values for first Lactation 305-d milk yield, fat
yield, and fat percentage in a Bos taurus xBos indicus multibreed dairy population
in Thailand. Thai Journal of Agricultural Science 35:339-360
Koonawootrittriron, S., M. A. Elzo, S. Tumwasorn, and K. Nithichai. 2002b. Estimation
of covariance components and prediction of additive genetic effects for first
lactation 305-d milk and fat yields in a Thai multibreed dairy population. Thai
Journal of Agricultural Science 35:245-258.
Koots, K. R., J. P. Gibson, C. Smith, and J. W. Wilton. 1994. Analyses of published
genetic parameter estimates for beef production traits. 1. Heritability. Anim. Breed.
Larsen, M., and P. Madsen. 1999. The CEBUS project: History and overview. Online.
Available: http://www. csc.fi/ttn/ccb99/articles/MLarsen.pdf. Accessed February
Lidauer, M., R. Emmerling, and E. A. Mantysaari. 2002. Accounting for heterogeneous
variance in a test-day model for j oint genetic evaluation of Austrian and German
Simmental cattle. Proceedings of the 7th World Congress on Genetics Applied to
Livestock Production. CD-ROM communication no 20-09.
Lidauer, M., and I. Stranden. 1999. Fast and flexible program for genetic evaluation in
dairy cattle. Online. Available: http://www. csc.fl/ttn/ccb99/articles/MLidauer.pdf.
Accessed Jan. 30, 2004.
Lee, C., C. P. Van Tassell, and E. J. Pollak. 1997. Estimation of genetic variance and
covariance components for weaning weight in Simmental cattle. J. Anim. Sci.
Manrique, C. 1992. Changes in genetic effects and variance components for preweaning
growth traits in an Angus-Brahman herd. PhD. Dissertation, University of Florida,
Martinez, V. G., V. E. Vega M., M. Montafio B., and J. Rosales A. 2002. Parametros
geneticos para crecimiento hasta los 12 meses de edad en bovines Charolais.
XXXVIII RNIP. Puebla, Mexico.
Mathur, P. K., B. Sullivan and J. Chesnais. 1998. A new method for assessing
connectedness between herds. Proc. National swine Improv. Fed. Conf. and annual
meeting. On line: http://www.nsif.com/Conferences/1998/mathrtm Accesed
Feb 29, 2004.
Meyer, K. 1989. Approximate accuracy of genetic evaluation under an animal model.
Livest. Prod. Sci. 21:87-100.
Meyer, K. 1997. DFREML version 3.a. Program to estimate variance components by
Restricted Maximum Likelihood using derivative-free algorithm. User notes.
Animal Genetics and Breeding Unit, University of New England, Armidale NSW
Meyer, K. 2002. "RRGIBB S" A program for simple random regression analyses via
Gibbs sampling. Proceedings of the 7th World Congress on Genetics Applied to
Livestock Production. CD-ROM communication no 28-27.
Meyer, K., and W. G. Hill. 1997. Estimation of genetic and phenotypic (co)variance
functions for longitudinal or "repeated" records by restricted maximum likelihood.
Livest. Prod. Sci. 47:185-200.
Misztal, I. 1993. RENUM-data preparation program for sire and animal models.
University of Illinois, Urbana, IL. p 7.
Misztal, I. 1999. Complex models, more data: simpler programming? Online. Available:
http://www. csc.fl/ttn/ccb99/articles/IMisztal .pdf. Accessed Jan. 30, 2004.
Mohiuddin, G. 1993. Estimates of genetic and phenotypic parameters of some
performance traits in beef cattle. Anim. Breed. Abstr. 61:495-522.
Mrode, R. A. 1996. Linear models for the prediction of animal breeding values. CAB
International. Biddles Ltd, Guilford. UK.
Northcutt, S. L., R. L. Willham and D. E. Wilson. 1991. Genetic parameters for nuclear
and nonnuclear inheritance in three synthetic lines of beef cattle differing in mature
size. J. Anim. Sci. 69:4745-4753.
Nufiez-Dominguez R., L. D. Van Vleck, and L. V. Cundiff. 1995. Prediction of genetic
values of sires for growth traits of crossbred cattle using a multivariate animal
model with heterogeneous variances. J. Anim. Sci. 73:2940-2950.
Olson, T. A., M. A. Elzo, M. Koger, W. T. Butts, Jr., and E. L. Adams. 1990. Direct and
maternal genetic effects due to the introduction of Bos taurus alleles into Brahman
cattle in Florida. I. Reproduction and calf survival. J. Anim. Sci. 68:317-323.
Pearson de Vaccaro, L. 1973. Some aspects of the performance of purebred and
crossbreed dairy cattle in the tropics. Part 1. Reproductive efficiency in females.
Pefia, R. E. 1998. Genetic characterization of criollo cattle in Bolivia. M Sc. dissertation.
Univ. of Florida. Gainesville.
Poyton, R. O., and J. E. McEwen. 1996. Crosstalk between nuclear and mitochondrial
genomes. Annu. Rev. Biochem. 65:563-607.
Quiroz-Valiente, J., V. Vega-Murillo, A. Rios-Utrera, and M. Montafio-Bermutdez. 1994.
Milk yield and composition of Brahman and Fl crossbred Angus, Hereford,
Charolais and Brown Swiss x Zebu cows. Proc. 5th World Congress on Genetics
Applied to Livestock Production 20:379-390.