t 5:. a K:
131?!“ u: 3.5
3:15 ... a .
:5; .

4.“.

5.5.8.
.. .. 2‘. 4
n )3 I 5!.

f .4. Uri}...

.1.
‘3’...
5.. . Kin... Jailiab
(ll)! ‘5’} ‘1 4
a i: 3:: purvﬂﬂnbn: :y v.5...
"I: in Vii-5.33.4.

 

3 :15... 311.1;
a D. 1...... 519355!
In; 6.3?!"

. l
3...: {1.39; 5:35.31
id "(3:53,
a . v
a. :

 

A . .35....astl .Ilir‘I
:.A!l..n.41rlt9 . .6: .) n. _ :-
IiuunnixL .

Iiiwﬁl. tier-414.: .r bl}...
#:1913313}

Q‘. .1}.

I:

 

. .5
31.4.. it

13135.1: . . ll! g..c:l..-r‘~.)ln!,.

l ga‘fu.}lotﬁn\lv1:;.

. . {as‘r11‘kﬁ‘$xas\§n§§

2.5.. .hv.\._.(n.q§.5).:.! I} .3...

. (‘51:)! I...
:.\ .531: It. A . I. I
. . 12!;

z ..y . A; 1r. :i
u...

1.21).!

..«I .: S .. ..

 

 

“’WWWW
._ , LIBRARY “

Michigan state
University

 

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11/00 cﬁlRODatoOuopGS—p I4

 

EFFECTIVENESS OF ALTERNATIVE BREEDING SCHEMES
FOR A DAIRY CATTLE POPULATION OF LIMITED SIZE

By

Chu-li Chang

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Animal Science

2000

ABSTRACT

EFFECTIVENESS OF ALTERNATIVE BREEDING SCHEMES
FOR A DAIRY CATTLE POPULATION OF LIMITED SIZE

BY

Chu—li Chang

There are many dairy cattle populations of approximately 100,000 cows in size in the
world that have sought genetic improvement. The geographically isolated dairy cattle
population in Taiwan was used to demonstrate alternative breeding schemes for their
rates of genetic improvement in production and their economical efﬁciency. Since the
source of germplasm import was from the U.S., the ﬁrst study was to estimate genetic
responses in individual traits, yield merit functions, and total merit ﬁmctions when selection
was based on criteria that were currently available to the US. dairy cattle breeders.

To maximize genetic improvement in milk yield, several practical breeding schemes
were designed in the second study. Each of the alternative breeding schemes focused on
the generation of genetically superior bulls that are to be use to breed cows in the
population, and each involved the use of semen imported from the US. These breeding
schemes belonged to three categories. The ﬁrst had breeding bulls produced by using
local genetic resources but their sires were foreign. The second used imported
germplasms to produce breeding bulls, which would not go through progeny test. The

third category included each of the schemes in the ﬁrst two categories plus a supplement

of additional imported germplasms. Ranking of cumulated genetic progress after 25 years
showed that the best breeding scheme was the one which used imported embryos to
produce bulls that would be screened by the performance of their patemal-half-sibs
before used for artiﬁcial insemination purposes. This scheme included a supplement of
5% imported pregnant heifers.

Economic efﬁciency of the same breeding schemes was assessed in the third study,
since in practice, a mere consideration of their ability to maximize genetic progress
would be insufﬁcient. In this study, the appropriate governmental agencies would be
expected to bear all costs for the infrastructure and operation of a breeding scheme.
Therefore, the cost factors of a scheme would include only those that would incur to
farmers such as costs of semen, artiﬁcial insemination, importing heifers, and the
increased costs of feeds and health and fertility problems due to increased milk
production. The economic efﬁciency of a breeding scheme was calculated as Net Present
Value of accumulated beneﬁt over a 25 years time horizon. The ranking of schemes by
economic efficiency was different from that by rate of genetic progress. However, the

exact same breeding scheme was deemed to be the best according to both criteria.

ACKNOWLEDGEMENTS

First, I would like to express my sincere appreciation to my major advisor and
mentor, Dr. Ivan. L. Mao. He has enabled me to organize my ﬁeld of study with his
constructive criticism and excellent guidance. I also wish to express my appreciation
especially to Mrs. Cecilia Mao for her spiritual support, encouragement and often care of
my study life while I was studying at MSU.

I am grateful to my academic guidance committee members, Drs. Dennis Banks, Ted
Ferris, Richard Pursley, and Carl Ramm for their assistance and constructive contribution
to my work. Gratitude is extended to Dr. Chein Tai, former Director General of Taiwan
Livestock Research Institute and also my guidance committee member, and Dr. Mao-
chiang Chen, Director of Hsin—chu Dairy Branch of TLR1. They made my study at MSU
possible by securing a 3-year fellowship from Council of Agriculture, Executive Yuan,
and by allowing me to retain my position and salary for the same period.

I thank all of my contemporary members in the Animal Breeding & Genetics Group
in the Animal Science Department at MSU in particular Dr. Rob Tempelrnan, Dr. Joe
Zhang, Dr. David Norris, Dr. Morten Rye, Thomas Mark, Lars Nielsen, and my ﬁ'iends
Sarah Wang, Jackie Ying, and staff members at Hsin-chu Dairy Branch of TLR1 for their
help and friendship.

To my wife Chui-lan and my parents and brothers, thank you all for your support,

encouragement, and conﬁdence.

TABLE OF CONTENT

LIST OF TABLES ........................................................................................................ vi

LIST OF FIGURES ....................................................................................................... xi

INTRODUCTION ......................................................................................................... 1
CHAPTER 1

LITERATURE REVIEW ............................................................................................... 5

INTRODUCTION ................................................................................................ 5

TRAITS FOR GENETIC IMPROVEMENT IN DAIRY CATTLE ....................... 7

Yield of Milk and Milk Components ........................................................... 7

Body Conformation Traits ........................................................................... 8

Health ......................................................................................................... 8

Longevity .................................................................................................. 11

GENETIC EVALUATION ................................................................................. 12

Current USDA Animal Model ................................................................... 12

Current Selection Criteria .......................................................................... l7

BREEDING STRATEGIES ............................................................................... 19

Artiﬁcial Insemination .............................................................................. 20

Multiple Ovulation and Embryo Transfer .................................................. 23

Marker Assisted Selection .......................................................................... 26

Genotype-Environment Interaction ............................................................ 27

Strategies Based on Importation of Genetic Material ................................. 28

Computerized Model to Study Breeding Schemes ..................................... 30

GENETIC RESPONSE FROM SELECTION ..................................................... 31

Discrete Generations and Overlapping Generation .................................... 33

Gene Flow Procedure ................................................................................ 34

The Time-Dependent Population Inventory Approach ............................... 3S

Realized Genetic Gain ............................................................................... 35

Reasons for Realized Genetic Gain Less Than Theoretical Gain ................ 36

ECONOMIC EFFICIENCY OF A BREEDING SCHEME ................................. 39

The Beneﬁts of Genetic Improvement ....................................................... 39

The Total Costs of Breeding Scheme ........................................................ 40

Evaluation Criteria of Economic Efﬁciency .............................................. 41

DAIRY CATTLE INDUSTRY IN TAIWAN ..................................................... 43

Background ............................................................................................... 43

Dairy Industry ........................................................................................... 43

Milk Production Costs and Beneﬁt of Milk ............................................... 44

Dairy Farming ........................................................................................... 44

Breeding Related Programs and Practices ................................................. 46
CHAPTER 2
EXPECTED RESPONSES FROM SELECTION ON CURRENTLY AVAILABLE
CRITERIA OF GENETIC MERITS. ............................................................................ 49
ABSTRACT ....................................................................................................... 49
INTRODUCTION .............................................................................................. 50
MATERIALS AND METHODS ........................................................................ 54
Traits and Indexed Merit Function ............................................................ 54
Calculation of Genetic Responses and Correlated Responses .................... 55
Genetic Parameters Used In Calculating Genetic Response ....................... 56
RESULTS AND DISCUSSIONS ....................................................................... 59
Genetic and Correlated Responses in Single-trait to All Selection Criteria 59
Total Genetic Responses in the Indexed Merits ......................................... 65
CONCLUSIONS AND IMPLICATIONS ........................................................... 66
CHAPTER 3
PRACTICAL BREEDING SCHEMES TO MAXIMIZE GENETIC IMPROVEMENT
IN A DAIRY CATTLE POPULATION OF LIMITED SIZE ....................................... 68
ABSTRACT ....................................................................................................... 68
INTRODUCTION .............................................................................................. 69
MATERIALS and METHODS ........................................................................... 72
Descriptions and Assumptions of the Target Population ............................ 72
Assumptions for the Imported Gerrnplasm ................................................. 72
Current Breeding Scheme (CBS) ............................................................... 74
Alternative Breeding Schemes .................................................................. 75
The Rate of Genetic Change under a Breeding Scheme ............................. 84
Genotype by Environment Interaction between Populations ....................... 91
RESULTS AND DISCUSSIONS ...................................................................... 92
Estimated Genetic Trend for the Exporting Country .................................. 92
Estimated Genetic Levels and Genetic Progress ......................................... 93
Comparison among Proposed Breeding Schemes ...................................... 95
The Effect of Genotype by Environment Interaction ............................... 102
CONCLUSIONS AND IMPLICATIONS ......................................................... 104

vi

CHAPTER 4
ECONOMIC EFFICIENCY OF ALTERNATIVE BREEDING SCHEMES

FOR A DAIRY CATTLE POPULATION OF LIMITED SIZE .................................. 107
ABSTRACT .................................................................................................... 107
INTRODUCTION ............................................................................................ 108
METERALS AND METHODS ........................................................................ 109

Current Breeding Scheme and Proposed Alternative Breeding Schemes .. 109

The Model of Evaluation- NPV ............................................................... 112

Calculating NPV of Beneﬁt ..................................................................... 114

RESULTS AND DISCUSSIONS ..................................................................... 117
IMPLICATIONS AND CONCLUSIONS ......................................................... 124
CONCLUSIONS ........................................................................................................ 127
APPENDIX ................................................................................................................ 131
I. Appendix A ................................................................................................. 131

II. Appendix B .................................................................................................. 140
BIBLIOGRAPHY ....................................................................................................... 147

vii

LIST OF TABLES

Table 2-1. Literature estimates of heritability and variances for the three yield traits (kg),

15 type traits (points), somatic cell score (scores) and productive life (mo) ................... 57
Table 2-2. Literature estimates of genetic correlations between traits ............................. 58
Table 2-3. Literature estimates of phenotypic correlations between traits ....................... 59

Table 24. Genetic responses in single traits to selection from single-trait and multiple-
trait indices ................................................................................................................... 60

Table 2-5. Correlated genetic responses in body frame traits by different selection criteria
...................................................................................................................................... 62

Table 2-6. Correlated genetic responses in udder traits by different selection criteria ..... 63

Table 2-7. Correlated genetic responses in the indexed merits by different selection
criteria ........................................................................................................................... 65

Table 3-1. Descriptions and assumptions of the dairy cattle population ......................... 73

Table 3-2(a). Abbreviation and deﬁnition of breeding schemes using local bull-dams .. 79

Table 3-2(b). Abbreviation and deﬁnition of alternative breeding schemes using imported

germplasm and their features ......................................................................................... 80
Table 3-3. The selection intensity in each selection path for breeding schemes .............. 85
Table 3-4. The accuracy of selection in each selection path for breeding schemes ........ 86

Table 3-5. The generation interval in each selection path for breeding schemes ............. 88

viii

Table 3-6. Assumed genetic trend in the US dairy populations and genetic merit of
gemplasm imported from the US into target population ................................................. 93

Table 3-7. Estimated annual genetic progress of breeding stock and the target cow
population for the current breeding scheme (CBS) ....................................................... 94

Table 3-8. Estimated genetic progress in the nucleus population for the MOET/AI/N S .94

Table 3-9. The estimated genetic progress of target population at year 25 under different
breeding schemes .......................................................................................................... 97

Table 3-10. Genetic improvement in milk yield (kg) at year 25 for the alternative
breeding schemes with supplement of F S40 or F820 to breed 30% or 50% cows ........... 98

Table 3-11. The genetic change in each selection path for breeding schemes ............... 101

Table 3-12. The effect of different genotype by environment interaction ( rg ) on genetic
improvement of milk yield for the proposed breeding schemes at year 25 ................... 104

Table 4-1. The total costs of milk production and beneﬁt per hundred kg for target
population dairy farmers in 1997 ................................................................................. 115

Table 4-2. Annual breeding costs to farmers for the proposed breeding schemes ......... 116
Table 4-3. Economic evaluation for individual scheme at year 25 ............................... 119

Table 4-4. Economic evaluation for individual scheme with supplement of FS40 or F8
20 to breed 30% or 50% cow population at year 25 ..................................................... 122

Table 4-5. The effect of genotype by environment interaction on NPV of relative beneﬁt
to that of CBS from the proposed breeding schemes at year 25 .................................... 123

LIST OF FIGURES

Figure 3-1. Estimate trend of genetic progress of the target cow population under
proposed breeding schemes over 25 years .................................................................... 95

INTRODUCTION

There are many dairy cattle populations of limited size in the world that operate
independently, either geographically or politically. The number of cows in these limited
size populations is varied. But a typical size of the population is about 100,000 cows. Each
population has their own unique features in management system, feeds and feeding, and
climatic conditions that would inﬂuence the level of milk production. Each population
also has made efforts to improve the efﬁciency of milk production by breeding, feeding,
and health care. In order to improve milk yield by selective breeding, much money has been
spent to import germplasm. The dairy cattle population of 100,000 cows in Taiwan is a
typical such population with a breeding scheme currently has 20% of cows bred by local
untested AI bulls, 20% bred by natural service bulls and 60% bred by foreign semen.

The primary breeding goal for dairy producers is to improve milk yield. Other traits
such as fat and protein yields, type ﬁnal score, linear functional type traits, somatic cell
count, calving ease, and productive life may be also important depending on breeding goals.
Any of these traits must ﬁrst be recorded so to have animals recorded and their relatives
genetically evaluated before it can be considered in brwding goals. For the goal of increase
proﬁtability and efﬁciency of a cow, a joint selection on a collection of traits should be
considered. However, which traits would be included with yield traits in selection
program and how much each trait should be emphasized are debatable issues in every
dairy breeding industry. The relative importance given to these traits was different from

breeder to breeder and done in different manners.

More than 30 selection criteria are available from USDA and from Holstein
Association, USA. Among these criteria, there are three kinds of criteria for genetic
selection: (1) PTA (Predicted Transmitting Ability) values for selecting a trait of
particular interest, (2) Product Value Indexes, and (3) Total Merit Indexes. For single trait
selection, PTA value could be used as selection criterion. If producers prefer to seek
improvement in several traits simultaneously, Product Value Index and Total Merit Index
are available. These selection criteria were developed for various selection objectives and
management conditions from farm to farm. Information on genetic responses or correlated
genetic responses of all economic traits and indexed merits to various selection criteria were
limited. All of these responses were useful to the breeders as a guide to set up their breeding
program, and to the AI organization in choosing bulls’ dam and in sampling young bulls. It
is important to understand the genetic responses and correlated genetic responses in single
traits and merits, when selection was based on a variety of criteria that were currently
available to the US. dairy cattle breeders.

There is a variety of breeding schemes taken by dairy developed countries, resulting
in large genetic progress in production levels. A breeding scheme is a practice of genetic
selection with application of certain reproductive technology or biotechnology. The goal
of selection is to allow animals with high breeding values to be parents of the next
generation. The ﬁrst and most important reproductive technology that had applied to
breeding scheme in dairy cattle is Artiﬁcial Insemination (A1) with frozen semen.
Artiﬁcial Insemination and Progeny Testing (AIPT) have been the backbone of dairy
cattle improvement in developed dairy countries, and can result in rates of genetic gain of

2 to 3%/yr (Van Vleck, 1986). Embryo transfer (ET) technology enables females to have

increased family sizes and shortened the generation intervals. The use of Multiple
Ovulation and Embryo Transfer (MOET) in a nucleus to produce bulls for genetic
improvement of dairy cattle herd could reduce generation interval and increase genetic
gain by 4 to 5 % which would be superior to those achieved in conventional breeding
schemes (Dekkers and Shook, 1990). Other technologies applied to breeding schemes
such as large factorial mating breeding designs with use of in vitro embryo production
(IVEP) and the use of genetic marker as aids to select animals which was called Marker
Assisted Selection (MAS) is also studied by several geneticists. (Brascamp et al.,1993;
Kashi et al., 1990; Mackinnon and Georges, 1997; Ruane and Clooeau, 1996; Spelman
and Gairick, 1997).

To bring genetic material of superior merit from a foreign population into a local
population is one rapid way to make genetic progress. Germplasms imported can include
young bulls, pregnant heifers, frozen semen, and frozen embryos.

Most of the populations of limited size are developing populations, which have not
applied the breeding schemes used in developed countries because environment
conditions and biological and economical resources are not the same. Theoretical
effectiveness of a breeding scheme in its rate of improvement is altered in practice due to
economics and success rate of the reproduction technology involved. Many studies
evaluated the cost effectiveness of breeding schemes (Hill, 1971; Dekkers and Shook.
1990)

At present time, Taiwan has a fairly well developed breeding and recording
infrastructure in place; i.e. technologies such as A1, embryo transfer and DHI. The dairy

cow population approximates 100,000 in size which will likely stay at the same

considering the environment and resources for dairy farming. Yet, there does not exist a
long-term breeding policy for the dairy population, which would accelerate genetic
progress in milk yield efﬁciently both biologically and economically. The dairy
population in Taiwan of 100,000-cow-population will be used as target population.

The overall goal of this work is to identify a breeding scheme that is optimum in both
the rate of genetic progress in milk production and the economic efﬁciency of
application. The speciﬁc objectives of this thesis were:

(1). To calculate the expected genetic responses in individual traits and indexed
merits from selection based on selection criteria currently available in the US. This was
done because all designed breeding schemes involved U.S.-imported germplasms.

(2) To design alternative breeding schemes that can be applied realistically in the
target population, and to compare their expected rate of genetic progress.

(3) To compare the same schemes for their economic efﬁciency with respect to the

conditions and biological and economical resources in application.

CHAPTER 1

LITERATURE REVIEW

INTRODUCTION

The dairy population around the world is decreasing in size. This fact means the
dairy producers must become as efﬁcient and as proﬁt-minded as possible. There are
many dairy cattle populations of limited size in the world that operate independently,
either geographically or politically. The number of cows in these limited size populations
might be varied. But a typical size is about 100,000 cows. Each population has own unique
features such as management system, feeds and feeding, and climatic conditions that
would inﬂuence the level of milk production. Each population also has made efforts to
improve the efﬁciency of milk production by breeding, feeding, and health care. The
primary breeding goal has been to improve milk yield and has spent much money to import
foreign germplasm to improve milk yield. The geographically isolated dairy cattle
population of 100,000 cows in Taiwan is a typical limited size population.

The main goal of a dairy producer's genetic improvement program should be to
produce replacement cows with the greatest possible genetic capability for making a
proﬁt. Fulﬁlling this goal requires strong, healthy cows that produce high levels of milk
of desirable composition. Also, these cows must be able to stand the stress of high
production through many lactation with a minimum of special care.

To set up breeding goals, traits for improvement should be set up. Can these and other

traits be selected for or not depending entirely on if they are recorded and genetically

evaluated. Single—trait or multi-traits selection would result in correlated responses for
traits with genetic associations. Because correlated response may have economic
“consequence, the impact of selection for milk on the total economic merit of cows should
be monitored. It is useful for the breeders and producers to know all of these responses as a
guide to set up breeding program.

A breeding scheme is a practice of genetic selection with the application of speciﬁc
reproductive technology or biotechnology. The goal of selection is to allow animals with
high breeding values to be parents of the next generation. Artiﬁcial Insemination and
Progeny Testing (AIPT) have been the backbone of dairy cattle breeding scheme in the
developed dairy countries. Other reproductive and biotechnology's applied to breeding
schemes, such as large factorial mating breeding designs with use of in vitro embryo
production (IVEP) and the use of genetic markers as aids to select animals which was
called Marker Assisted Selection (MAS), have also been studying by several geneticists.
To bring in genetic material with higher genetic merit into a local population is another
way to make genetic improvement especially when the foreign population is genetically
superior to the local population.

Breeding schemes involve different costs, both in the initial capital and in annual
rtmning expenditure. The monetary returns from genetic improvement accumulate over a
long period of time, and the pattern of returns may be erratic in early years. It is
necessary to know the returns and costs for any breeding schemes so that sound
investment decisions can be made.

The purpose of this chapter was to review the literature regarding the responses in

various genetic merits from several known selection criteria, evaluation system to

establish selection criteria, and the dairy cattle breeding strategies in the developed
countries. Then, in the following chapters, several breeding schemes would be designed
in order to seek an optimum breeding strategy for the speciﬁc conditions and resources in

Taiwan.

TRAITS AND MERITS FOR GENETIC IMPROVEMENT

Sale of milk is the major source of income for most dairy producers. Milk yield,
therefore, should receive heavy emphasis in selection programs. Other traits such as
stress resistant traits (rear leg set), fertility trait (rump angular), and mastitis resistance
(SCS) are also important. These traits are associated with proﬁtability and efﬁciency of a
dairy cow as well as the producer's net return.

Yield of milk and milk components

The largest single source of costs in milk production is management overhead costs
of keeping a cow. The more milk a cow produces, the more these overhead costs are
spread out and the less the cost per kilogram milk is produced. Thus, the number of cows
with high genetic potential for milk yield should be maximized. With continued breeding
for higher genetic output, a farmer will produce the same milk with fewer cows and at
substantial saving in costs.

Yet, farmers get paid for milk volume with an adjustment for fat and protein
percentage. This misleads some farmers to focus their attention on fat and protein
percentage. In fact, farmers get the payment from price for the milk volume excluding the
fat and protein yield and from the kilograms of fat and protein they produce. So the

payment system is also really aimed at fat and protein yield. If selection is on fat

percentage or protein percentage, it will actually tend to decrease milk production
because the genetic correlation between fat percentage, protein percentage and milk yield
is negative. So, to maximize farmers' income from milk production, select sires for
improving herd should be high genetic evaluations for yield traits, not for fat and protein
percentage.

Body conformation traits

Type traits have long been considered a relatively important trait of dairy cattle. In
the early days of dairy cattle breeding, a cow's outward appearance (type ﬁnal score) was
probably the best indicator of her producing ability. Extreme dairy character, capacious
barrel, and good udder were suggestive of high production. Other type traits were
considered important because of their perceived positive relationship with longevity.
Cows with a strong front end, that were wide, a little slope rump, with structurally sound
feet and legs, and well-attached udder were seemingly more trouble-free could be
recorded early in a cow's lifetime, they have some value for predicting herd life.

Many dairy producers think that single-trait selection for milk may be detrimental to
cows, particularly for traits related to strength, stamina, and survival. Numerous studies
(Everett et al., 1976; Foster et al., 1988; Norman and Van Vleck, 1972; Sieber et al.,
1987) had investigated relationships of milk yield and conformation. Genetic correlation
of milk yield and ﬁnal score was near 0.1. For individual type traits, dairy character has
consistently had favorable relationships with milk yield (Foster et al., 1988; Norman and
Van Vleck, 1972). Udder depth has been negatively associated with milk yield (Foster et
a1 1988). Some studies (Foster et al. 1988, Sieber et al., 1987) found unfavorable

relationships between milk yield and udder attachments. Estimates of genetic correlation

of milk yield with most other type traits have usually suggested little meaningful
association. Meland et a1. (1982) reported no clear relationship of milk yield and type
when milk selection and control lines were compared.
Health
1. Mastitis and Somatic cell counts

Somatic cell counts (SCC) from a day's milk are the best indicator of the extent to
which the gland is involved in ﬁghting a mastitis infection. The DHI program provides a
monthly SCC which identiﬁes those cows with high SCC, which might be sub-clinical
mastitis. Sub-clinical milk appears to be normal. Yet, milk yield is depressing, and
composition may be altered. Sub-clinical mastitis may become clinical. Genetic
correlations between SCC and mastitis are moderately high (Emanuelson et al., 1988).
An SCS is a logz transformation of somatic cell count and is related to mastitis incidence.
The transformation makes the values nearly normally distributed and has the simple
interpretation that each increase of 1 is a doubling of the somatic cell count. As with
yield, lactation averages for up to the ﬁrst ﬁve lactations are included. Somatic cell score
(SCS) evaluations were implemented in 1994 in the United States (Schutz, 1994). SCS
evaluations are centered on breed average, which is 3.2 for Holsteins. Strandberg and
Shook (1989) demonstrated that selection to reduce the rate of increase in mastitis that
accompanies selection for milk yield could be economically prudent. Rogers et al (1991)
suggested an index that combines linear type traits with production and SCC.
2. Reproduction

Calving ease (CE) is of economic important trait in dairy cattle and should be

considered in breeding programs. Dystocia is a reproductive problem of dairy cows,

especially for ﬁrst-calf heifers. The economic costs of dystocia include loss of calf,
veterinary fees, farmer labor costs, increased risk of subsequent health and fertility
problems, increased culling, and reduced production. Factors affecting calving case can
be separated into maternal and fetal (or direct) components. Maternal calving ease (MCE)
refers to characteristics of the dam giving birth (e.g., pelvic dimensions). Direct calving
ease (DCE) refers to characteristics of the calf (e.g., calf size) (Meijering, 1984).

Although dystocia can be reduced by proper management procedures, such as heifer
rearing and feeding during gestation, selection and breeding strategies have been
identiﬁed as important additional tools to reduce dystocia in the short and long term
(Meijering, 1984). Many studies have identiﬁed small but signiﬁcant genetic components
to both DCE and MCE; heritabilities ranged from .03 to .20 (Cue and Hayes, 1985,
Meijerin, 1984; Weller and Gianola, 1989.). Genetic correlations between CE and
production traits was close to zero (Meijering, 1984). Signiﬁcant genetic relationships
between MCE and conformation traits had been found, especially conformation traits
associated with pelvic shape and pelvic dimension (Dadati. et al., 1985).

The National Association of Animal Breeders funds the analysis of calving difﬁculty.
Using a categorical model (Berger, 1994), calving ease evaluations are calculated for
bulls at Iowa State University. Evaluations are reported as the expected percentage of
difﬁcult births for ﬁrst-calf heifers giving birth to a bull calf during the winter. These
evaluations have been particularly helpful in promoting the use of Al dairy bulls with
virgin heifers.

Reproductive performance is one of the major factors inﬂuencing the proﬁtability of

a dairy herd. Reproductive performance affects the amount of milk, breeding costs, rates

10

of voluntary and involuntary culling, and the rate of genetic progress for traits of
economic importance. Reproductive efﬁciency of bulls is usually measured by non-
return rate, which is commonly deﬁned as the proportion of cows that were inseminated
and did not return for another service within a speciﬁed number of days, such as 28 or 56
days or commonly 60 to 90 days. The non-return rates can be used to derive useful
measures of the reproductive efﬁciency of a bull (Koops et al., 1995).

Longevity

. Longevity has been deﬁned in many ways. Ducrocq et al (1988) discussed two
deﬁnitions of PL. True PL characterized the aptitude for a cow to remain in the herd
given the management's criteria for selection and culling. Functional PL represented the
ability to delay involuntary culling. Genetic selection for increased PL is expected to
result in improved general health, production, and reproduction. They can increase dairy
farm proﬁtability by decreasing the number of replacement heifers needed, by allowing
rearing costs to be dispersed over a longer period, and by increasing the number of cows
producing at a mature level (Allaire and Gibson.,1992; Rogers et al., 1989).

Productive life (PL) (VanRaden and Klaaskate, 1993) in the United States is deﬁned
as the number of months in milk (with a maximum of 10 months per lactation) until the
cow is 84 months old. PL evaluations are the same as for yield evaluations (VanRaden
and Wiggans, 1995). To add accuracy for bulls, information on PL from type traits is
added (Weigel eta1.,l996). The Holstein Association USA combines evaluations for PL,
milk and fat yields, and linear type traits to calculate an approximate multi-trait
evaluation. This enhancement currently is applied only for Holstein bulls and has its

greatest beneﬁt for bulls with a type evaluation but that have a yield evaluation that is

11

based on daughters that are too young to contribute PL information.

GENETIC EVALUATION

Advances in computing power and reduction in cost per unit of computing resource
had made possible continuous improvement in evaluation system. Computers have
allowed more and more complete models and statistical procedures to be used for sire and
cow evaluations (Lee et al., 1985). Improvement in genetic evaluation system to
accurately rank animals for selection is the most important practice in dairy cattle
breeding. Accuracy of identifying superior sires and dams relies on best-unbiased
prediction of breeding values. Genetic evaluation system has been frequently modiﬁed to
get best-unbiased prediction of breeding values, exploiting advances in computer
technology and evaluation methodology.
Current USDA animal model

The current USDA animal model can be represented as:

y=Mm+Za+ZAgg+Pp+Cc+e

Where y represents standardized milk, fat, or protein yield. m, a, g, p, and C are
vectors of effects for management group, random portion of additive genetic merit,
unknown-parent group, permanent environment, and herd-sire interaction, respectively.
M, Z, ZAg P and C are incidence matrices for these effects; and e is error. Predicted
Transmitting Ability (PTA) is the estimate of genetic superiority for bulls and cows by
using animal model (VanRaden and Wiggans, 1991).

The current animal model, which was implemented in 1989, has been modiﬁed

frequently to meet the needs of the US. dairy industry and to exploit advances in

12

computer technology and evaluation methodology. The predicted transmitting ability
(PTA) reported to the dairy industry is half the animal's breeding value.

Genetic evaluations of linear type traits for Holsteins are calculated by the Holstein
Association USA. A multi-trait animal model (Misztal et al., 1993) is applied with
adjustment for heterogeneous variances (Weigel and Lawlor, 1994). A canonical
transformation is used to create uncorrelated traits for analysis. Because the data include
more than one scoring per cow, a permanent environmental effect is included.

Misztal and Wiggans (1988) obtained more precise measures of the accuracy of
individual evaluations using an iterative procedure that was computationally affordable
but lacked easy interpretation. Methods to summarize accuracy of PTA provided by the
additional sources of information included in evaluations also were needed. The total
amount of information provided by records of the animal and all its relatives is
summarized by REL (Reliability) through a simple function of total DE (Daughter
Equivalent) from parents, own yield, and progeny adjusted for mates. A DE is the amount
of information about a parent's genetic merit available from a single lactation of a single
daughter in a herd. Daughter equivalents can be converted to Reliability of a PTA for
milk using the formula REL = n / (n + 14) where n is the number of daughter equivalents.
REL ranges from 0 (least accurate) to 99 (most accurate).

In practice, REL for females seldom exceed 80 percent. For most bulls, milk, fat,
and protein will have the highest reliabilities because more daughters have performance
records on those traits. Reliabilities for type traits tend to be lower since not all daughters
are scored. Heritability affects reliability and, as a result, different type traits have

different reliabilities. Low heritability will cause somatic cell score (SCS) reliabilities to

13

be lower than for milk, even if all daughters have SCS data. Heritability of productive life
(PL) is the lowest of all traits with PTA's calculated on a national basis. Heritability is
less than 10% for PL expressed in mature daughters and less than 5% when predicted
from type data on young cows (V anRaden and Wiggans,]991; Schutz, 1994).
INTERBULL

Due to increasing trade in semen, embryos and livestock in the world, breeders want
to make accurate and fair comparisons between animals performing both within and
across countries. One way to compare the evaluation between two countries are
conversion equations. But they were used on the scales of both countries for a number of
bulls and permitted estimates of genetic merit from one country to be expressed on the
genetic scale of another country. Conversion equations were calculated only between
pairs of countries and force all converted evaluations to be on the prediction line.
Because converted evaluations were re-scaled original evaluations, bulls with converted
evaluations were not re-ranked on the new scale (Wiggans etal.,1992).

Because of differences in genetic evaluation methods, differences in genetic base,
differences in breeding goals, and differences in environment among countries, it is
difﬁcult to compare genetic merit evaluated by each country. An organization called
INTERBULL Center was established in 1991, under contract with the Swedish
University of Agricultural Sciences in Uppsala, Sweden. INTERBULL coordinated
international communication. They provided international leadership in researching and
developed methods for generating international genetic evaluations. Currently, the
International Genetic Evaluation Service provided by INTERBULL evaluates sires from

over 20 countries for three milk production traits - fat, protein and milk yield. Other traits

l4

such as health and ﬁtness traits, including SCS and productive life will be included in the
ﬁtture.

INTERBULL uses a scientiﬁcally advanced method known as Multiple Across
Country Evaluation (MACE) to calculate International Genetic Evaluations. MACE
combines information from each country using all known relationships between animals,
both within and across populations. It accounts for the possibility of animals re-ranking
between certain countries. This occurs when animals perform better in certain
environments than they do in others. Genetic merits are expressed in different ways and
units. Legitimate units are pounds of PTA for the United States, kilograms of PTA for the
United Kingdom, Ireland, and Israel, relative breeding values for Scandinavian countries,
and kilograms of EBV for all other countries that participate in INTERBULL
evaluations.

In 1997, INTERBULL had 34 member countries, with more likely to join in the
future. 54,000 Holstein Friesian sires from around the world received international
genetic evaluations expressed in PTA lbs. on the US base, in EBV kg on the Dutch base,
in EBV kg on the French base and so on. A bull with an EBV for milk of 1000 kg has an
EBV of 2200 pounds would be improper and unethical. Updated evaluations from
INTERBULL are made available in February and August on release This provides the
advantage of individual countries being able to identify those animals from around the
world which will perform best under their own unique farming conditions (Banos and
Sigurdsson. 1996). The Center will increase the frequency of its evaluations to four per
year to match the evaluation schedules of many of the major participating countries.

Now, INTERBULL and most national evaluations are calculated four times each year.

Although the equations will change with each subsequent evaluation, the changes usually
are expected to be small. However, when a country changes the deﬁnition of its genetic
base population, as occurs with the ﬁrst evaluation of each year in Canada and with the
mid-year evaluation in France, knowing the evaluation date becomes even more
important because changes in conversion equations can be quite dramatic when bases
change.

The best option always is to use the national or INTERBULL evaluations on a single
scale. The preferred way to compare the genetic merit of bulls from different countries is
to have the evaluations all expressed on the same scale. Fortunately, most bulls of interest
have International Bull Evaluation Service evaluations for milk, fat, and protein yields.
The ﬁle of ofﬁcial evaluations on the US. scale is available from the Animal
Improvement Programs Laboratory's (AIPL's) Internet web site. The conversion
equations are to convert from one scale to another for bulls without INTERBULL
evaluations for milk, fat, and protein yields. There are conversion equations developed by
INTERBULL to convert evaluations from other countries' scales to an US. basis. They
are available at the AIPL web site. For example, a bull that has a Netherlands estimated
breeding value (EBV) of 1700 kg for milk yield would be expected to be similar to an
US. PTA of about 2000 pounds. This conversion is only approximate the actual PTA
milk from INTERBULL. The conversion equations are directional, which means they
only can approximate an US. PTA from an EBV ﬁ'om other countries, but they are not
valid for estimating EBV's of other countries from an US. PTA or from EBV from other

countries.

Current selection criteria

1. Current selection criteria for single-trait in US.

PTA (Predicted Transmitting Ability) could be used as selection criteria for single-
traits of interest including milk, protein, fat, type traits, SCS and PL in dairy cattle. The
use of PTA's have two purposes: to rank animals for genetic merit and to estimate genetic
differences between animals. A cow with a PTA of 1500 for milk is expected to produce
daughters averaging 500 lbs. / lactation higher production as mature cows than daughters
of a cow with a PTA of 1000. The cow with a 1500 lb. PTA would rank higher than the
cow with the PTA of 1000 lb. Only animals of the same breed can be compared.

For single trait selection, PTA value (Predicted Transmitting Ability) could be used as
selection criterion. This value has been available for many traits of interest for genetic
selection. Some breeders apply selection pressure on only one trait.

PTA for Somatic Cell Scores (SCS) were published since 1993 in the US. The
heritability of SCS is approximately 9 percent. It will be more difﬁcult to improve the
SCS of the herd through selection than to improve milk production. The major way to
alter the Somatic Cell Count of a herd is through management changes, not through
genetic selection. If one selects on SCS, sires should be selected ﬁrst on production traits
and then sires that have above average SCS PTAs should be eliminated. In other words,
use PTA SCS as a culling criterion once production yields meet requirement. This
procedure does not decrease emphasis on production traits and will still let a producer use

a group of low SCS sires.

2. Current selection criteria for multiple traits in US.
(1). Production or yield indices
MF$ = $.10170(PTA M)+$.58(PTA F)
= 5(PTA M/625)+1(PTA F/22)
Relative emphasis on component traits 5: 1
MFP$ = $0.031 (PTA M)+$0.8 (PTA F)+$ 2.00 (PTA P)
= 2.67(PTA M/625)+1(PTAF/22)+1.7(PTA P/19)
Relative emphasis on component traits 3 : l : 2
CY$ = -$.00220 (PTA M)+$1.98(PTA F)+$1.716(PTA P)
= -(PTA M/625)+32(PTAF/22)+24(PTA p/l9)
Relative emphasis on component trait -1: 32 : 24
(Cheese market for Ayrshires, BS, Hs, M85 and R&Ws)
MF$, MFPS, CY$ are Product Value Indexes.
(2). Total merit indices
NMS = 0.7(MFP$)+$11.30 (PTA PL)-$28.22 (PTA SCS-3.2)
= 10 (PTA MFP$/71)+4(PTA PL/13.2)-1(PTA SCC/1.9)
Relative emphasis on component traits 10: 4: -1
(U SDA-DHIA Active AI bull evaluation.1995).
TPI = [$7.89(PTA P)+$2.22(PTA F)+$71.43(PTA T) +$12.18(UD)+
$6.5(FU) +$6.5(TP)+$6.5(RUH)+$4.87(RUW)+$4.06(UC)
= [3(PTAP/ 19.0)+1(PTAF/22.5)+1(PTAT/. 7)+. 65(UDC/. 8)
+. 35(FLC/. 85)] 50+576

Relative emphasis on component traits 3: 1: 1: 1:65: .35

18

Where .UDC=. 30(UD)+. 16(FU)+. 16(TP)+. 16 (RUH)+. 12(RUW)+ .10(UC)

Relative emphasis on component traits .3: .16216: .16: .12: .1

F LC=. 5(Linear traits of foot and leg)+ .5(Feet and Leg Score)

Relative emphasis on component traits 1: 1

Linear traits of foot and leg = . 48(FA)+. 37(RLRV)-. 15(RLSV)

(Holstein Association, 1999)

Originally, these indexes were derived based on the current milk pricing system. In
1994, standard milk price was $.12/lb of milk, reﬂecting a $12.00/cwt price for milk
containing 3.5 percent fat and 3.2 percent protein. The above formula is speciﬁc for
Holsteins because 3.2 is the breed average SCS for Holsteins However, the price paid for
milk components is expected to change as consumer demand changes over time. If proﬁt
is considered as the long-term objective, it is then important to consider selection indexes
using future milk-pricing system because breeding decisions are made approximately 3
yr. before the eventual replacements begin to milk.

The TPI includes type traits that impact proﬁtability of the animal. This index heavily
emphasizes PTA protein with no direct weight on PTA milk. However, selection for fat

and protein yields tends to increase milk yield as well.

BREEDING STRATEGIES
A breeding scheme is a practice of genetic selection with application of certain
reproductive technology or biotechnology. To design breeding strategy is to develop

strategy with respect to selection, mating, and population structure that optimize the

19

relationships among these factors such that genetic improvement was maximized or
optimized.

Artiﬁcial Insemination

(1) Progeny Testing

The Artiﬁcial Insemination and Progeny Testing (AIPT) scheme identiﬁes bulls with
superior genetic merit for milk, fat, protein and type through their relatives and many
daughters' performance. In progeny testing, the young bulls are usually chosen on
pedigree information for production and type and on their own conformation. Then each
young test bull was mated to a number of cows; a number large enough to ensure that
each bull had a sufﬁcient number of daughters for a reliable proof. A sampling procedure
is to identify those bulls receiving a favorable sample of alleles ﬁ'om parent.

After young bulls sampling, one method is to lay off the young bulls while waiting
for their proofs (the lay-off policy). During the lay-off period, little semen was collected
regularly, frozen and stored. When the proofs become available, low proof bulls were
culled along with any semen which had been stored and the selected bulls were placed in
the proven stud for regular service (Hinks, 1970). The second method is the collect &
slaughter policy. Large amounts of semen were stored from testing bulls which wee
slaughtered. When proofs were available, semen from the undesirable bulls was disposed
(Johannsson, 1970). The two methods gave different rates of genetic progress and
differed in running costs. The cost of storing the required volume of semen for
approximately four years amounted to only about 25% of that of keeping the bulls for six
years while waiting for their progeny test results (Johannsson, 1970).

Progeny testing schemes were expensive to run especially for small size population.

20

Progeny testing scheme required that good infrastructure be in place; i.e. good road
services, a milk recording scheme, an registry system, Al services etc. Fewson (1989)
demonstrated that economic beneﬁts of testing might be low for small populations. He
calculated proﬁts of progeny testing for different population sizes. The proﬁts for a
large population (500,000) were almost the same as those for the medium population
(100,000 cows) and those for the small population (20,000) were almost half those of the
large population. Hunt et al (1974) found that for a population size of 115,000 cows, the
progeny group sizes gave the highest annual genetic gains when 50% of the cow
population was milk recorded and 40% of the tested population were bred to young bulls.
The additional beneﬁt of increasing the progeny group size from 20 to 30 daughters per
bull was very small. Lindhe (1968) found that the accuracy of progeny testing increases
only slightly if the size of the progeny groups exceeded 60 daughters.
(2) Selection of dairy bulls on half-sib records

Selection on the basis of half-sib performance has the disadvantage that the
correlation between the average of half-sibs and the breeding value of the candidate was
only half the correlation between the average of the progeny and the breeding value of
the candidate. But the advantage was that the age of bulls at selection was reduced by 2
years (approximately by one-third). Owen (1989) described a half-sister selection of
young bulls scheme as an alternative to the progeny testing for the genetic improvement
of milk yield in a large population of dairy cows. The results indicate that the half-sib
testing scheme would lead to an eventual rate ‘of genetic progress very similar to the
progeny test scheme. However, the annual cost of the half-sib testing scheme was less

than a progeny test scheme, because the waiting time for candidate bulls was completely

21

eliminated. In addition, improvement of the dairy cows showed up sooner in the new
scheme by the 10th year from the start. The cumulative genetic improvement was four
times greater than for progeny testing.

(3) Untested AI Young Sires

Potentially superior young animals needed to be identiﬁed early. Pedigree Indexes
(P1) or Parent Average (PA) had been utilized for this purpose for many years. Pedigree
Indexes are .5 and .25 of the genetic evaluation of sire and maternal grandsire (MGS)
respectively. PA's could be used to rank and select young bulls. There is no better
predictor of future milk production on young animals than PA. A bull's PA was
calculated as one-half of his sire's Predicted Transmitting Ability (PTA) plus one-half of
his dam's PTA for a particular trait. Although PA was a good predictor of the bull's
eventual proof, some young sires will come through with proofs below their parent
average, while others will come through with higher proofs. The net result was that the
average proofs for a group of bulls was usually quite close to the overall average of the
bulls' PA (Powell and Norman,l989).

One of the weaknesses of the PA had been the accuracy of predicted transmitting
abilities (PTA) for bull-dams. Though the implementation of the animal model greatly
increased the accuracy of cows PTA, there was still a tendency of overestimating ETA
for bull-dams (Ferris et a1. 1991, Graham et al, 1991, Mao et al, 1991). Zhang et al.
(1994) investigated that indexing young bulls for AI were compared relative to
subsequent progeny tests on the same bulls. Correlations among the pedigree indexes and
eventual proofs for traits in the Lifetime Proﬁt Index ranged from .6 to .8. Of the top

40% of young bulls on the basis of pedigree indexes, 86% (93 out of 108) were also in

22

the top 10% on actual proofs. No bull with P1 in the bottom 30% had an actual proof that
reached the top 10%. So, the key to use semen from untested young bulls should not mate
a large portion of the herd to a few young sires (Powell and Norman,l989).

Thompson and Freeman (1972) found that the use of untested sons of progeny tested
bulls gave the most rapid genetic gain in small populations. Hunt et a1. (1974) compared
estimates of genetic gain from progeny testing programs with estimates when the same
population was bred only to young bulls. The all young bull studs gave similar rates of
genetic progress to progeny testing schemes for the smallest cow population.

Multiple Ovulation and Embryo Transfer

Where PTAI were neither available nor inadequate, bulls for the commercial
population could be produced from a MOET nucleus scheme. Success of such a scheme
would depend on the successful use of the advanced technology. MOET offered a chance
to improve selection intensities in the female pathway. Compared to progeny testing, the
accuracy of evaluation was reduced since the number of sibs was usually lower than the
number of progeny. For small populations, however, the differences in accuracy of
evaluation may not be large since the number of progeny per sire for test bulls could also
be quite low. Moreover, accuracy in evaluation would improve as success rates with
MOET improve. With MOET, bull selection could be made to predict a bull's breeding
value by the information on all collateral relatives (both paternal and maternal full sibs
and half sibs). The generation interval could be reduced when using sister records instead
of progeny records resulting in expected increased genetic gains. There were many
different MOET breeding schemes to select superior bulls. Ruane (1988) reviewed these

schemes and the rates of genetic change possible. MOET nucleus schemes offered the

23

highest rates of progress of all the schemes that use MOET. Adult MOET nucleus
schemes were the more feasible ones to adopt (Ruane, 1988).

MOET nucleus schemes were sensitive to embryo transfer success rates. Qualiﬁed
and experienced personnel and good recipient herd health programs were therefore a
must. Otherwise a proportion of cows fail to respond to super-ovulation and yield no
viable embryos or fewer embryos than expected. In Canada, the proportion of cows
selected that fail to respond was 20% (Lohuis et al., 1990). This would reduce female
selection differentials. When superior cows that failed to respond were replaced with
cows of lower genetic merit, lower rates of progress were achieved. Low success rates
could also result in a smaller number of families giving reduced accuracy of evaluation
for the bulls. Therefore, more donors than required should be selected. Embryo yields
per cow per ﬂush in the ﬁeld average about 7 (Lohuis et al.,l990), although yields of 8
embryos per cow per ﬂush were possible with experienced technicians(Hasler et al.,
1987). Pregnancy rates of 66% were possible with non-surgical transfers and will
increase with experience and practice (Rowe et al., 1980). Rates of 65% were reported
for the British MOET nucleus scheme (McGuirk, 1989) and 45% for ﬁeld MOET
schemes in Canada, respectively (Lohuis etal., 1990).

Another area of concern with MOET nucleus schemes in a limited size was the rate
of inbreeding and the result of reduction in genetic variation. This would result in lower
rates of genetic change. Inbreeding depression would have an affect on both ﬁtness and
production and reproductive performance (Falconer,l989). Inbreeding was important
with MOET schemes since the nucleus herds were likely to be small. The smaller the

number of donors, the lower the expected rate of progress, and the higher the risks

24

associated with inbreeding (Ruane and Thompson, 1991). The number of bulls required
for the commercial population and the level of inbreeding to be tolerated determined the
herd size. In a small nucleus herd, inbreeding would be limited if the initial base
population was of a wide genetic background, if factorial mating designs suggested by
Woolliams (1989) were used, and if the nucleus was kept open to introducing foreign
germplasm of genetic material from either a foreign population or the local elite
population. McCuirk (1989) suggested that a smaller herd was acceptable only if it was
possible to introduce outside animals of high genetic merit to the herd after the program
had commenced. High inbreeding rates were also due to the use of family information,
which increased the frequency of selection of sibs. Juga and Maki-Tanila (1987) and
Ruane and Thompson (1991) showed genetic response in milk production varied from
1.0% to 1.26% per year for the alternatives studied. These rates were substantially lower
than the 1.78% to 2.24% per year predicted for the Adult MOET nucleus schemes in
Nicholas and Smith (1983).

Fewson (1989) derived the optimum number of donor cows for different cow
population sizes. For a population size of 100,000 cows, he recommended 50 to 100
donor cows be selected. Dekkers and Shook (1990) did a ﬁnancial analysis of a breeding
company running a progeny test scheme and adopting a MOET nucleus scheme. They
found that closed adult nucleus schemes were competitive to the conventional progeny
test scheme only for large nucleus herds.

Dekkers (1992) reported conventional AIPT schemes could be considered as
dispersed open nucleus breeding schemes in which the nucleus consists of daughters

resulting from bull-darn matings. With selection on animal model BLUP EBV across age

25

groups, use of MOET on bull-dams for the production of young bulls was then
theoretically as effective in capitalizing on MOET with regard to genetic improvement as
utilization of formal MOET nucleus breeding schemes. The advantages of establishing a
separate nucleus station over a dispersed (open) nucleus herd mainly pertain to logistics
and to the quality and quantity of station versus ﬁeld information, where quantity refers
to the amount of information on sibs and the number of traits evaluated.
Marker Assisted Selection

In current animal breeding schemes, prediction of genetic differences between
animals was based on phenotypic observations, which depend on genetic and
environmental factors. Restriction fragment length polymorphism, variable number of
tandem repeats, and polymerase chain reaction made it possible to identify genetic
differences directly at the DNA level. These differences, which were called genetic
markers, were not likely to be quantitative trait loci (QTL) themselves, but they may be
linked to QTL (Soller, 1978). In a breeding scheme, use of phenotypic and marker data
could provide more information than phenotypes alone. When marker data were included
in the selection criteria, this process being referred to as marker-assisted selection (MAS),
accuracy of selection would be increased. Marker data could be collected early in life,
allowing selection at an early age. In this respect, MAS is similar to juvenile indicator
traits. The use of genetic markers in animal breeding involves four steps: 1)The search
for genetic markers; 2) establishment of a linkage map of the markers; 3) detection of
associations between markers and QTL; and 4) use of marker-QTL associations in the
breeding program (Woolliams and Smith, 1988).

Several quantitative trait loci (QTL) for milk production traits have been identiﬁed in

26

dairy cattle (Spelman et al., 1997; Vilkki et al., 1997). QTL could be utilized in marker-
assisted selection (MAS) breeding schemes. MAS for dairy cattle had been evaluated in
many studies and have shown that the rate of genetic gain could be increased by the
implementation of MAS (Brascamp et al., 1993; Kashi et al., 1990, Mackinnon and
Georges, 1997; Ruane and Clooeau, 1996; Spelman and Gairick, 1997).
Genotype-environment interaction

The introduction of foreign germplasm into another country was possible existence
of genotype by environment interaction. A trait measured in two different environments
is to be regarded not as one trait but as two. If the genetic correlation is high, then
performance in two different environments represents very nearly the same trait,
determined by very nearly the same set of genes. If it is low, then the trait is to a great
extent different, and high performance requires a different set of genes (Robertson,
1959b; Dickerson, 1962; Yamada, 1962). Such interactions had been shown to be
unimportant among temperate countries (Garabano et a1. 1989; Sieber and Powell, 1989).
Falconer (1989) proposed that the genetic correlation between the expression of the same
genotype in two different environments could be used as a measure of genotype by
environment interaction. The product-moment correlation between breeding values of
sires estimated in different environments was not inﬂuenced by differences in variation.
This was the method used in several studies (e.g. Petersen, 1975; Danell, 1982). When
genotype by environment interactions exist, the best genotype in one environment was
not the best in another. If the genetic correlation between countries was low, breeding
strategies based on local selection programs may be superior to those based on imports of

genetic material. For moderate genetic correlation, strategies based on imports could be

27

adopted. The response in the importing country should be estimated as correlated
response. The results reported for GXE interactions between temperate and tropical
countries were inconsistent and inconclusive possibly due to limited sets of data
available. In McDowell et a1. (1976) the genetic correlation estimates for milk
production in Mexico and the USA was high (.86) whereas in Abubakar et al. (1987), it
was moderate (.51). The lowest genetic correlation in performance between countries
found in the literature was .08, between Sri Lanka and Denmark (Buvanendran and
Petersen, 1980). Chauhan (1983) reported the genetic correlation between Denmark and
India was .11 while that between USA and India was .7.

INTERBULL estimated of genetic correlations among countries. Estimates would
normally range from .86 to .89 between two Northern Hemisphere countries and from .75
to .78 between a Northern and a Southern Hemisphere country. The model of national
evaluation was also taken into consideration: countries with similar national evaluation
models were assigned higher genetic correlation estimates.

Strategies based on importation of genetic material

The primary objective of importation of germplasm was to bring in genetic material
of higher genetic merit from foreign population and thereby raise the average genetic
mean of the importing population. The genetic impact depended on 1) How often
imported material was made; 2) the imported material was used for SB or SC; 3) the
difference in genetic means between the importing and the exporting country; 4) the
importance of genotype by environment interaction. Progress was likely to be fast when
the exporting country was much superior, when a large proportion of cows were mated to

foreign bulls, when imports were continuous rather than only once, and when genotype

28

by environment was not important. The main disadvantage of import strategies was cost
of imported germplasm. The importance of genotype by environment interaction
between importing countries and exporting countries had to be determined. If it was not
important, then strategies based on imports of genetic material or those combining
imports with local selection could be adopted (Smith, 1988).

For a developing country, semen was imported to breed commercial cows. Therefore,
the semen need not be from the best bulls in the exporting country but the bulls should be
of higher genetic merit than average bulls in the importing country. The price of semen
from such imported bulls should be competitive. Usually, semen importation programs
have always been combined with selection program.

Economic returns on farm level for semen importation were calculated for Colombia,
Mexico and Venezuela (Holrnann et al., 1990). Average economic returns were negative
for the three countries. This was because production levels in the three countries were
low so the authors recommended semen importation only in high producing herds and
when conception rates were high. Since production conditions and production levels
were favorable in importing countries, it may therefore be worthwhile to investigate this
strategy.

Holmann et a1. (1990) found that local processing of semen collected in Colombia
from imported bulls was better in economic terms than importing semen. A group of
young bulls could also be generated from imported embryos, produced by mating the best
cows and bulls in the exporting country. Young bulls from imported embryos were used
untested or some could go through the progeny-testing scheme. Population parameters for

the importing countries dairy population were within the range for which this strategy

29

was suitable. Using ET to produce young bulls for progeny testing has been estimated to
increase genetic gains by up to 15 percent (Petersen and Hansen, 1977, Hill and Land,
1976) and ﬁnancial gains by over 40% (Dekkers and Shook, 1990).

Computerized model to study breeding schemes

A model for studying breeding scheme would use mathematics, statistics, and
computer techniques to create the possibility of more extensive experimentation, describe
the phenomenon in questions, bring about relations in conditions and approaching those
to the real world so as to obtain a fuller understanding of a breeding scheme. So, the
model would provide information that may be too expensive or too time consuming to
obtain by experimentation. The application of model to animal breeding or even livestock
science and production research has increased greatly in recent years. Four reasons could
be forwarded to explain this phenomenon. In dairy cattle breeding schemes, genetic
improvement is dynamic in nature and inﬂuenced by various ﬁxed and random effects
and based upon biological productivity such as yield, the factor of fertility, reproductive
capacity, and mortality. Genetic improvement in a dairy population is accumulated over
time. So a computerized model is suitable for the study of this topic to obtain a more
complete understanding in the real dairy cattle breeding world.

The types of model given by France and Thomley (1984) could be divided into
empirical and mechanistic models, static and dynamic models, and deterministic and
stochastic models. An empirical model sets out principally to describe, whereas a
mechanistic model attempts to give a description with understanding. A static model is a
model that does not contain time as a variable while a dynamic model contains the time

variable explicitly. A deterministic model is one that makes deﬁnite predictions for

30

quantities (such as genetic progress) without any associated probability distribution. A
stochastic model, on the other hand, contains some random elements or probability
distributions within the model. Stochastic, semi-stochastic and deterministic models had
been suitable and used to study genetic improvement in long-term breeding experiments
(e.g. Kislev and Rabiner, 1979; Dekkers and Shook, 1990; Woolliams, 1989; Dekkers et
al.,l996). Stochastic models allow a more realistic representation of the process of
selection on a quantitative character. They also allow calculation of the variation in
response associated with breeding programs. However, stochastic simulations required a
large number of replicates. When stochastic model simulated several strategies and long
term predictions for a large population, they can be time consuming. Deterministic
models predict the results of given breeding schemes using formulae that were derived
theoretically and provide an understanding of the underlying parameters. Because of their
lower computational requirements, a large number of alternatives breeding schemes can
be evaluated. Therefore, deterministic models were selected for this study of several

breeding strategies.

GENETIC RESPONSE FROM SELECTION
Robertson and Rendel (1950) developed an equation for predicting genetic change
based on theoretical considerations. The equation was deterministic in nature to predict
expected genetic change for a deﬁned selection criterion, usually interpreted as genetic
value for a single trait. Symbolically, the equation for expected superiority is

AG=i"'r,,"'(Sa

31

Where AG : expected genetic response

1': selection intensity

rp : accuracy of genetic evaluation , p is the top fraction selected

based on genetic merit

6.: additive genetic standard deviation for milk yield.

Some conditions underlying the equation were normality, truncation selection for the
selection criterion, genetic gain at equilibrium and additive genetic variation constant
over time. The equation shows the superiority in average breeding value of selected
animals over those available for selection. Genetic improvement per year depends on the
generation interval, or the average time in years between when an animal is born and
when its offspring is born.

These factors can be used to determine which scheme would be expected to result in
the largest gain per year. For most traits, additive genetic value makes up most of the
genetic value and is the part, which is transmitted directly to progeny. Thus, selection
will be assumed to be for additive genetic value.

Rendel and Robertson (1950) predicted genetic change based on four selection
pathways: sires to breed bulls (SB), sire to breed cows (SC), dams to breed bulls (DB)

and dams to breed cows (DC). The estimator of annual genetic change (Ag) was given

as:

ZAG
A =———
g 2L

Where AG is the genetic superiority for each of the 4 pathways and L was the
generation interval for each of the 4 pathways. Sires of Bulls (SB) were used by Al
organizations to produce young sires for sampling. The SB contributes the greatest to

genetic progress because they were highly selected. Accuracy of selection was also high

32

because all SB were progeny tested and accuracy usually is over .9. Dams of Sires (DB)
were highly selected from over the top 2% of all cows in the population. Accuracy of
selection was less than the SS path, ranges from .5 to .65. Sires of Cows (SC) were
chosen by farmers to breed their cows. Intensity of selection was high but less than for
SS. In general, about one of ﬁve young sires was returned to heavy service, and 13% of
the tested population was used for progeny testing young sires. Dams of Cows were
chosen by dairy farmers to leave offspring in the population. The DC path contributes the
least to genetic progress because of low selection intensity (Everett, 1984). Korver and
Renkema (1979) found that the rate of genetic progress was hardly affected by culling
females for production. Therefore, the cow to breed cows pathway was assumed no effect
on genetic progress for all strategies.

Most progress was made in selecting sires to produce the next generation of bulls,
with 43% of the total genetic progress in this path. The second most inﬂuential factor in
genetic progress comes from selecting dams of future bulls (33%). Selecting a group of
sires to breed cows accounts for 18% of the total genetic progress, whereas the selection
of dams to produce cows was the least important, accounting for only 6% of the total
(Rendel and Robertson, 1950).

Discrete generations and overlapping generation

Discrete generation occur when all of the offspring are kept till the last-born was
mature. Selection was then made and the selected individuals were all mated at more or
less the same time. The generation interval was the interval between the matings made in
successive generations (F alconar, 1989).

In dairy cattle populations, the generations are not discrete but were overlapping.

33

When generations overlap, the selected offspring were mated as soon as they were
mature. The generation interval could be calculated as the average age of the parents at
the birth of their selected offspring. This means that the individuals present at any time
were of different ages and at different stages of their life-cycles. Furthermore, individuals
differ in length of life and consequently in their opportunities for reproduction. The
longer-lived individuals had greater chance of contributing offspring to the next
generation than the short-lived do. Cows presented at any time are of different ages.
Selection of parents was more or less a continuous process. Because the genetic trend
exists, animals of different age classes should be regarded as coming from different
distributions of genetic values with different means and variance. Response from the
selection of parents could be calculated by multiplying the genetic superiority of parents
by the proportion of their genes present in later generations (Ducrocq and Quaas,l988).
Gene flow procedure

Hill (1974) presented gene ﬂow procedure using matrix formulation based on the
Rendel and Robertson (1950) formula. With this procedure, response from the selection
of parents was calculated by multiplying the genetic superiority of parents by the
proportion of their genes presented in later generations. Ruane and Smith (1989) used this
method to evaluate MOET schemes. The gene ﬂow model used was:

M(t) = P * [M(t-i) + AG (0]

where M(t) was a vector of average genetic merit of animals at time t, P was a gene
transmission matrix, M(t-i) was a vector of average genetic merit of animals at time (t-

i), and AG(t) was a vector of the genetic superiority of selected animals.

34

The time-dependent population inventory approach

Ducrocq and Quaas (1988) presented a time-dependent population inventory
approach for predicting response for a dynamic selection system. The approach would
ﬁrst identify newborn female groups that come from the same age group sires and the
same age group dams in the year. Then compute the with-in group selection superiority
and genetic merits of the selected parents who contribute genetic gains to the newborn.
The estimate of the annual genetic progress of the cow population in the year was the
weighted genetic merit mean of different newborn animal group.

Realized genetic gain

Estimates of genetic gain in dairy cow populations have not been equal to theoretical
gain. Realized genetic progress should be monitored to evaluate the success of breeding
programs (Van Vleck.,1977.,1986., Pearson.1984). Many studies have examined genetic
trend by regressing estimated breeding values on time ( Lee et al., 1985, Powell et al.,
1977) or regressing of production on time (Powell and Freeman. 1974). In those studies,
actual gain was considerably less than what was possible under ideal circumstances.

Van Tassel] and Van Vleck (1991) estimated average genetic selection differentials
for the four paths of selection for each year of birth. Selection differentials for paths of
sires of bulls, dams of bulls, sires of cows, and dams of cows averaged over all years
were 405, 395, 239, and 42 kg respectively, and for the most recent 5 years 884, 598,
235, and 28 kg respectively. Genetic trend based on the average selection differentials
and generation intervals would be 34.9 kg/yr, but based on the latest S-yr periods and
considering parents of grade cows, the genetic trend would be 57.2 kg/yr.

Breeding experiments have to maintain an unselected control population, which are

35

expensive to maintain because they quickly fall behind selected lines for productivity.
Legates and Myers (1972) reported direct and correlated responses to selection for milk
yield with control lines. Estimates of annual response for milk yield were 121 kg for
North Carolina Holsteins. Hickman (1971) found an annual trend of 4.5 kg for fat from
selection for solids yield of Holsteins. Boettcher et a1. (1993) reported a selection
experiment that was conducted over a 25-yr period to measure long-term responses to
selection of Al sires based on milk alone. The genetic trend from 1967 to 1988 for the
selection line was estimated to be approximately 125 kg/yr.

Reasons for realized genetic gain less than theoretical gain

1. Generation Intervals

Generation intervals were likely to be much longer than necessary, especially for
dam of bull and sire of bull paths (Westell.l984,.Lee et al., 1985). The sum of paths in Al
situations might be about 30 years, whereas the theoretical goal was 24 to 25 years. If
gain was predicted to be 100 kg per year, assuming a summed interval of 24 years, the
gain was reduced by 20% to 80 kg per year.

Van Tassel] and Van Vleck (1991) estimated generation intervals averaged for all
years by path were 10.2, 6.4, 9.3, and 5.1 yr. For the most recent 5 years, the interval
were 11.0, 6.4, 8.9, and 4.9 yr respectively. Estimates of annual trend were considerably
less than the potential rate of 96 kg/yr because of longer than necessary generation
intervals, and smaller selection differentials than theoretically possible.

2. Selection emphasis on other traits
A common practice among bull studs was to set minimum levels for several factors

including type and fat test that must be met before a cow qualiﬁes as a bull dam no matter

36

how good her genetic evaluation for production. Each such minimum standard, in theory,
would reduce the expected genetic superiority for production. These results indicate that
selection practices in that period were not optimum for maximum improvement of milk
production. Decreased emphasis for non-yield traits, while reducing generation intervals,
could dramatically improve genetic trend in the future (Van Vleck, 1986).
3. Problems with genetic evaluations

A major concern has been the failure of genetic evaluations of bull dams to predict
their sons evaluations as well as theory would predict (Vinson, 1984). The general result
has been that evaluations of dams based on other than ﬁrst lactation records hadn't
increased accuracy of predicting sons' proofs ( Murphy et al., 1982). In general, the use of
only ﬁrst lactation records in the genetic evaluation of the dam results in regressions of
son's or daughter's evaluation on sire's proof, dam's evaluation, and maternal grandsire's
proof. Such regression agreed closely with theoretical approximations of these regression
coefﬁcients. When the dam's evaluation was based on all lactation records, the regression
coefﬁcient for her evaluation was smaller than theory would predict. The regression
coefﬁcient for the maternal grandsire's proof seemed to compensate for the loss of value
of the dam's evaluation by changing from an expected small negative coefﬁcient to a
small positive coefﬁcient.
4. Inbreeding

Inbreeding could lead to a depression in phenotypic performance (inbreeding
depression) due to dominant gene effects. Keller et a1. (1990) found that the reductions in
genetic response due to inbreeding depression were m that ore severe than reductions due

to inbreeding effects on additive genetic variance alone. The rates of inbreeding were

37

high with selection in small populations. However, the rate of inbreeding was reduced by
introducing foreign unrelated animals into the breeding stock and was not important
when p0pulations were open to exchange of stock. Inbreeding would be important when
progeny testing in a closed population and in the nucleus herd for the MOET scheme.

The rate of inbreeding per year (F) was estimated using Wright's formula (1931):

FD=AF* D* he;p

Where F9 is inbreeding depression in a unit, D is the % depression per percent
increase in inbreeding coefficient.

5. Genetic variation (62,) and Bulmer effect

Genetic variation refers to the variability of additive genetic values within a
population for milk yield under selection. It relates to the difference between the average
individual and the top individuals. The more superior the top individuals are over the
average, the grater the potential to make genetic gain through selection.

Bulmer (1971) showed that the genetic variation was temporarily reduced by the
selection of parents and ancestors, but regenerates quickly when selection was relaxed.
This is because a group of selected parents represents one tail of the phenotypic
distribution, and in consequence their phenotypic variance must be less than that of the
whole population from which they were selected. So when calculating the genetic
responses, the temporary reduction in genetic variation due to selection of parents and
ancestors should be taken account (Falconer, 1989).

6. Other factors
Records that provided misinformation possibly because of misidentiﬁcation, or

preferential treatment, or non-random mating and treatment reduced genetic progress

38

primarily through the dam of bull and sire of bull paths.
Selection differentials were much smaller than thought possible for SC and SB,
which might be due to emphasis on traits other than milk production and high percentage

of sampled sires returned to service with little selection based on production traits.

ECONOMIC EFFICIENCY OF A BREEDING SCHEME

It is important to include all relevant costs and returns when evaluating the efﬁciency
of a breeding scheme. A sound investment of a breeding scheme should give high genetic
progress at reasonable costs and maximum beneﬁts to the dairy industry. Costs are
compared with returns to determine which scheme gives the greatest beneﬁt. The cost
and returns from genetic improvement for different breeding schemes accumulate over
time and are realized over different period. They must be expressed in a comparable unit.
The usual method of comparison is through discounting future costs and revenues to
current values. The net present value (NPV) can be used as an economic evaluation
criterion. In economical comparison of the expected beneﬁts from different breeding
schemes, the production of whole population of cows is of interest.
The benefits of genetic improvement

The major beneﬁt of genetic improvement. is the additional milk yield produced by
improved cows over that of cows from current breeding scheme. There are some
secondary beneﬁts due to increased milk production. For example, increased production
will inﬂuence all of the commercial dairy population as well as other production sectors

of the dairy industry. These include the increased imports of dairy genetic material, local

39

breeding infrastructure, and the growth of milk processing and feeds industries. All these
secondary beneﬁts are difficult to evaluate and are therefore excluded in this study.

The total costs of breeding scheme

1. Milk production costs:

Milk production costs are the costs incurred by the dairy herds. They include buying
the farm, building the dairy, fencing, electricity, labor, dairy equipment, tractor, feeds,
veterinary services, medicines, shipping, selling charges, electricity, insurance, and
miscellaneous costs. These costs and basic overhead expenses (e.g. living expenses of
farmers) are assume to remain the same no matter what scheme is adopted and are all
excluded ﬁ'om the analysis.

For a ﬁxed herd size, costs might change slightly with increased milk production
level. Breeding costs would increase because more semen is required for higher
producing cows. The milk costs for increased production due to genetic improvement
only considered feed costs and health costs in most studies (Ferris and Troyer, 1987;
Mpofu etal., 1993).

2. The investment costs:

These are the costs for setting up a breeding scheme and are incurred once at the
beginning of the program. However, If some items need replacement during the
operational period, the replacement costs for such items are also considered as investment
costs. Some items may outlast the evaluation period and will have a salvage value. Other
investment costs are fencing and building, electriﬁcation, roads, building a laboratory, lab
equipment, semen storage facilities and bull handing equipment. For semen import

schemes, the main investment costs are storage building and semen storage equipment.

40

3. The operation costs:

The operation costs are recurrent costs needed to run the scheme every year. For
most of schemes, the operation costs include bull procurement costs, bull maintenance
costs, and semen collection, storage and distribution costs.

The bull procurement costs are the rearing costs up to the time of purchase by bull
station. Bull maintenance costs per bulls per year include labor, feed, veterinary services
and medicines, stationery, and some miscellaneous costs. The operational costs increase
from year to year in the ﬁrst ﬁve years as the bull station reaches its full capacity.

For schemes involved in imported semen, the main operating costs are the cost of
semen and shipping, handling and export charges as well as storage and distribution
costs. Imported semen are priced according to their percentile group for PTA M. Other
running costs include semen storage costs.

For the MOET scheme, the operation costs are the costs of collecting and transferring
embryos, costs of collecting semen, administrative costs (to screening and selecting
donors) and high technical input. For the scheme involved in embryo import, operation
costs include the cost of embryos, costs of mating, shipping, and handling and export
charge as well as embryo transfer.

4. The contingency cost:

Contingency cost is the cost to allow for losses from damage during handling semen.
The number of semen or embryos imported is more 1.05 times than that required.
Evaluation criteria of economic efﬁciency

There are many criteria for evaluating breeding scheme economic efﬁciency. Net

Present Value (NPV) was the most commonly used (Petersen et al.,l974., Fox et

41

al.,l990., Dekkers and Shook.1990a).
The NPV is the net present value of the incremental income generated by the

investment. The following equation is used to calculate NPV:

r=r Br - Cr

NPV: -—-;
1:] (1‘1"!)

Where B. is the value of beneﬁts for year t, G the cost in year t, T is the period of
evaluation, and d is the discount rate. When the calculated value of BCR is less than 1
(i.e. the present value of costs at the given discount rate are higher than the present value
of the beneﬁts), it means expenditures and investment costs are not ﬁrlly recovered.

If the NPV is negative the present value of the beneﬁt stream is insufﬁcient to
recover costs. Therefore, independent projects that give a positive NPV are the ones to
be selected. The NPV is the preferred measure when the projects are mutually exclusive
(Gittinger, 1982) and can be used to rank such projects. Mutually exclusive projects are
projects of a nature that if one is chosen, the others cannot be undertaken. The strategies
to be evaluated in this study are mutually exclusive. The strategy giving the highest NPV
is the one to be selected. The problem with the NPV measure is that it can only be used

when the estimate of the discount rate is accurate.

42

DAIRY CATTLE INDUSTRY IN TAIWAN

Background

Taiwan is located on the Tropics of Cancer in the Paciﬁc Ocean east of China
mainland and consists of the main island Taiwan and the Pescador Matsu and Kimem
Island. The island Taiwan covers an area of 35,989 kmz. The population is almost 22
million. The climate is mainly subtropical and in the southern parts tropical. The average
temperature for the island is 28°C. Average temperatures are 150 to 200C in the winter
and 300 to 35°C in the summer. July is hottest month and February the coldest month,
with average temperature of 32°C and 16°C, respectively. The average relative humidity
is 85% with a variation from 70 to 98 %.
Dairy Industry

Taiwan is hardly an ideal place for dairy farming. Milk is produced only from
Holstein. Due to high temperature and high humidity, the main constrains that dairy
farmers have long been confronted are the cows' summer infertility syndrome, mastitis,
and foot and leg problems. Especially, the summer infertility inﬂuence the cows to
produce less milk when the price is up in summer and to produce more when the price is
down in winter. Because of the government's policy support and the increased consumer
demand for fresh milk, however, the dairy industry has been growing at some 16%
annually for the past years. Yet, the self-sufﬁciency of milk and milk products in total
still remained at 15-20% in recent years. A large variety of dairy products are imported

and sold at relatively low prices. To be competitive, almost all raw milk produced locally

43

is processed into ﬂuid milk products and recently some milk has been processed into
yogurt.
Milk production costs and beneﬁt of milk

Production costs varied from farms to farms and were inﬂuenced by herd size and
herd production levels. The costs of milk production include feed, labor, veterinary
expenses and medicines, tractor operating costs, selling charges, electricity, insurance,
transport costs (delivery of milk to depots) and some miscellaneous costs. The major
costs were feed (50% of total costs), labor (12%), veterinary expenses and medicines
(8%) and transport costs (8%). For feed costs, cows were fed according to production and
NRC requirements Labor costs include wages. Tractor operating costs cover fuel,
lubricants, repairs and maintenance. Veterinary, medicine and miscellaneous costs can be
grouped into one. This group are included dipping, semen price and AI fee, detergents for
the dairy equipment, etc. These were expressed as a percentage for one-kg milk
production.

Farmers were paid for the amount of milk produced. The basic price of milk is
adjusted for the hygienic and compositional quality of the milk delivered. In 1997, the
standard milk price was $US.62/kg of milk, containing 3.5 percent fat. In 1997, beneﬁt
from milk per kg were $NT4.55/kg ($USO.14/kg) obtained from the ﬁle of Council of
Agriculture.

Dairy farming

There are 843 dairy farms in Taiwan in 1997. The average herd size reached 113

heads with a range from 40 to 500 cows per herd, and the average land holding was 1.86

hectares. In order to reach an economic scale of 100-150 heads (total number), pregnant

44

heifers had been imported in the past ten years and this made the herd size increased
more rapidly recently. Most dairy farms belong to a family type with 2-3 persons. Due to
mechanization, they are able to manage a herd between 100 and 150 cows in 10 working-
hour per day.

There are two main local grasses used for dairy cattle, namely, Napier grass and
Pangola grass. Corn silage has been getting popular in recent years, because of its
competitive price. Alfalfa hay, in the form of bale and cubes, and bale Bermuda grass hay
have been importing from North American countries and used as supplement. A
considerable amount of local produced agriculture by-products such as peanut vine hay,
corn stover, tomato pomace, soybean pomace, brewers grains, sorghum distillers grains
are used in dairy rations. As for the concentrate, most com and soybean depend on
importation. Since 1985, computer feeding of concentrate had been introduced, which
made the frequently and accurately feeding of concentrate according to body weight,
parity, milk production, and fat percentage of milk. More recently, Total Mixed Ration
(TMR) system has been adopting by many dairy farmers. Farmers using TMR system
would save much labor, reduce cows' digestion disorder, get much more stable milk
composition, and increase milk production as well.

Being in a hot climate, dairy farmers built simple yet practical barns with both sides
open, except those near the seashore and in the northern area where a wall on one side is
needed. The new or remodeled barns usually have enough space for mechanical operation
and also they are high and have insulated roof for good ventilation. The free stall barns
are concrete ﬂoors with some bedding with rice hulls and sawdust. Most farms milk cows

in their milking parlor with pipeline milking machine. Yet a few small farms still use

45

bucket system to milk cows in the barn. To increase efﬁciency and save labor,
automation including automatic crowd gate, automatic milking recording, automatic teat
cups remover are gradually become popular.
Breeding related programs and practices
1. Dairy Herd Improvement (DHI) program

There were 230 dairy herds with about 13,000 milking cows enrolled in 1996. The
average herd size for milking cows was 54 cows with an average parity of 2.85; the
average milk yield (305-2X-ME) was 6,572 kg and the average daily milk yield was 20.3
kg. The lowest daily milk yield was in August (17.67 kg); the highest was in December
(21.06kg). The average percentage of milk fat and protein were 3.68% and 3.16%,
respectively. The average milk somatic cell count (SCC) was 60,080/m1. The cows were
calved for the ﬁrst time at an average age of 26 months with calving interval 15 months.
The average calving interval was 446 days and the average number of days dry before
calving was 110 days. Calving was generally throughout the year, although most calvings
were during the six months from October to April and the frequency of calving declines
from May to September. The average days open was 163 days, and the services per
conception were 2.7. The data indicated the reproduction efﬁciency must be enhanced,
and milk production will be upgraded if the SCC goes down.

As the current payment scheme for milk yield include a penalty or premium related
to the bulk tank SCC, farmers are being forced to participate in DHI in the next few years
and the goal is to have 50% farms on DHI.

2. Artiﬁcial Insemination (Al)

Al was introduced into Taiwan in 1970 and now is available to all dairy farmers

46

through extension service or done by farmers themselves. About 80% of cows are
artiﬁcially inseminated to frozen semen from either local untested bulls (20%) or foreign
progeny tested bulls (60%). Most of imported semen was from the US, and some from
Canada, Japan and Holland. Emphasis was put on genetic merit for milk production and
price. There is a trend that selection pressure for the SC pathway is increasing. Price of
imported semen ranges from US$12 per straw to $US 50 per straw while semen from
local bulls is $US6.

A governmental bull station was established to select young bulls, to collect semen
and sell to farmers. Because only small proportion of cow population was on DHI, bulls
raised at the station were not progeny tested. But they are from mating which are of high
genetic merit of parents. These untested bulls provided semen for AI service to mate
about 20 % of cows. The farmers would like to use semen from these untested bulls
because of their higher conception rate, cheaper price and more or less contribution to the
genetic progress.

About 20 % of cows were settled by natural service bulls to promote the reproductive
efﬁciency, especially during the hot season from June to October. Some cows were sired
from imported semen with their PTA M from the top 10% of foreign progeny tested
active bulls with 81% REL.

3. Embryo transfer

Embryo transfer technology (ET) was introduced into Taiwan in 1980 and the
importation of frozen embryos was introduced in 1985. Success rates of ET were 40% for
frozen imported embryos and 60% for fresh embryos produced from local cows (Lee,

1993). The primary reason for embryo imports in Taiwan is to produce superior bulls and

47

high genetic merit females for production. Importing embryos instead of bulls are more
efﬁcient due to easy transportation and less restrictive health requirements. Animals born
from local recipients and raised under local conditions should be more resistant to the
local diseases.

4. Culling and importation of live animals

A high involuntary culling rate and a low reproductive efﬁciency results in no
selection for dams of cows and cause very little population expansion. The main reasons
for involuntary culled cows are reproductive sterility, mastitis, and foot problems. The
percentage for each of reasons was 33%, 21%, and15% respectively. Low reproductive
efﬁciency is due to 2.7 A1 services per conception resulting in a calving interval of 16
month.

Since 1986, farmers who wanted to expend their herd size imported heifers from
foreign countries. Pregnant or yearling Holstein heifers were brought into Taiwan, mainly
from the North American and some from Canada, Australia, New Zealand and Japan.
From 1985 tol992, importation of pregnant heifers of 5~10% as cow replacement per

year was practiced from the US, Canada, and Japan.

48

CHAPTER 2

EXPECTED RESPONSES FROM SELECTION ON CURRENTLY AVAILABLE
CRITERIA OF GENETIC MERITS

ABSTRACT

Genetic responses in 20 single traits and in ﬁve indexed merits were estimated, when
selection was based on a variety of criteria that were currently available to the US. dairy
cattle breeders. Single trait selection criteria were PTA values for milk, fat, protein, type,
somatic cell score (SCS), and production life (PL). Yield merit criteria were MFS, MFPS,
CY$ and total merit were NMS, and TPI. Comparisons were made after one round of
selection in a discrete generation with ﬁxed selection intensity and accuracy. The greatest
response in milk yield was from direct selection on PTA of milk, but responses ﬁ'om
selection on MFS, MFP$ and W8 were only slightly less. Increase in fat yield was, as
expected, the greatest when selection was directly on PTA of fat, but the increase in protein
yield was the greatest ﬁ'om CY$. The greatest response in type could only be achieved by
selecting PTA of type and TPI. Selections on most criteria would lead to an increase in SCS
with the least increase from selection on PTA of fat and NM$, but selections involving type
resulted in large negative responses. All selection criteria would lead to favorable responses
in linear type traits with a few exceptions, including fore udder attachment, udder depth and
teat placement. The greatest return in any merit was obtained by selection directly on that
speciﬁc merit index, but selection on PTA of yield traits and other merit indices would lead
to almost as much response. Selection on PTA of type would result in very trivial positive

responses in merits.

49

INTRODUCTION

Milk products are one of the most important food products for human consumption in
the world. Dairy farming is worldwide and much work has been done to improve efﬁciency
of milk production by breeding, feeding, and health care. Primary brwding goal has been to
improve milk yield. Other traits such as fat and protein yields, type ﬁnal score, linear
functional type traits, somatic cell count, calving ease, milking speed, and productive life
may be also important depending on breeding goals. Whether or not these and other traits
can be selected for depends entirely on if they are recorded and genetically evaluated.

Mastitis was the most costly health problem of dairy cows. Many studies were
conducted on improving udder health and reducing mastitis. SCC in milk was recorded on
most DHIA programs and might serve as an indicator of mastitis in breeding program
(Renea. 1986). As distribution of SCC is not normal, and its relationship with milk yield was
not linear (Shock, 1982), Dairy Records Processing Centers transforms SCC to somatic cell
score (SCS), which is the log; transformation of SCC. Somatic Cell Score (SCS) has been
accepted by the National Cooperative DHI program as a standard recording scale for SCC.
SCS had been used for genetic evaluation and has become an important economic trait of
dairy cattle. Genetic selection on SCS has been considered as one approach to improve
resistance to mastitis in future generation (Shook, 1989; Shock and Schutz, 1994). Rates of
improvement through genetic selection were likely to be slow but the cost involved in
genetic enhancement of disease resistance was small compared with the large cost of
treating clinical mastitis and the milk yield lost from sub-clinic mastitis (Roger, 1993;
Standberg and Shook,1989.).

Productive life (PL) was the trait used to estimate genetic merit for longevity. This trait

50

could be thought of as the ability of a cow to avoid culling. Increased productive life
contributes to the proﬁtability of dairy yield by decreasing replacement costs and by
increasing the percentage of cows producing at mature levels. The productive life of dairy
cows average about 3.5 yr. (Norman. et al., 1981) in US, which was much less than their
biological potential. Burnside et al. (1984) concluded that milk yield was economically
twice as important as herd life. Due to low heritability and PL is expressed at a later age
than the other traits. Many researchers have tudied the importance of productive life in dairy
selection programs. Although study methods differed, ﬁndings were quite consistent.
VanRaden and Klasskate (1993) suggested a method to evaluate PL from completed and
predicted data and made PL evaluation available. PL had higher economic value than many
traits currently evaluated and might be one-third as important as yield, based on an average
of recent estimates (Allaire and Gibson, 1992; Dekkers.1993.; Harris and ﬁeeman,1993;
Van Arendonk.1991).

One of the purposes of the linear type classiﬁcation program was to identify and to
emphasize traits associated with longevity and mastitis resistance. Feet and legs have always
been an important issue in dairy cattle breeding. Several studies have been conducted to put
on feet and leg traits and their correlation with stayability (McDaneil, 1995; Uribe et al.,
1995). Many studies have also been conducted to ﬁnd out the relationship between linear
udder traits and SCS (Seykora and McDaniel, 1986; Rogers,1990; Schutz and
VanRaden,1993). They concluded that udder depth, fore udder attachment, teat placement
had relationships with SCS.

Currently, dairy farmers in US. has more than 30 selection criteria available ﬁ'om

genetic evaluations done at USDA and other sources. Among these criteria, there were three

51

kinds of criteria for genetic selection: 1) PTA values for selecting individual traits, 2)
Product Value Indexes for selection of a composite of production traits, and 3) Total Merit
Indexes for selection of a collection of traits deemed to be economically important.

For single trait selection, PTA value (Predicted Transmitting Ability) has been available
for many traits of interest for genetic selection. Single trait selection criteria deal with only
single trait improvement. It is for the goal of increase proﬁtability and eﬂiciency of a cow
that a joint selection on a collection of traits is considered. However, which traits would be
included with yield traits in the selection program and how much each trait should be
emphasized are debatable issues in every dairy breeding industry. The relative importance
given to these traits is different from breeder to breeder and done in different manners.

For multi-trait or merit selection, a selection index (Hazel, 1943) was used to determined
appropriate weights for the traits included in a selection criterion. Merit selection criteria
allows to improve several yield traits and non-yield traits simultaneously. Among efforts in
the U.S., Rogers and McDaniel (1989) applied the selection index method by utilizing
involuntary culling in an aggregate genotype to represent the composite costs of health,
reproduction, and husbandry attributes. Traits considered in the index were milk yield, udder
depth, teat placement, and foot angle. Rogers (1993) used information on SCS, udder depth,
teat placement, and foot angle along with milk yield to identify optimal sire indexes. Misztal
(1992) used a restricted index to calculate maximum response in milk yield and maintain
udder depth at its current value. Computer software for designing selection criteria, which is
also getting popular, allows breeders to place selection emphasis on the traits which they
consider to have the most economic beneﬁt in their herd (Cassell, 1986)

That selection on one trait would lead to genetic change in another trait was called

52

correlated response. The correlated response depends on the genetic correlation between the
two traits. In fact, when selection was based on one index, each of the component traits in
the index would have a correlated response as well as other traits not in the index. There
would also be a correlated response in another indexed merit. Genetic improvement is
permanent and accumulates over time. Information on genetic responses or correlated
genetic responses of all economic traits and indexed merits to various selection criteria were
limited. All of these responses were useful to know to the breeders for guidance to set up
their breeding program. It would be useful to the AI organization in choosing bulls’ dam and
in sampling young bulls. If there would be undesirable responses for some economic traits
on a selection criterion, and if the breeder cannot tolerate some of these increased
undesirable responses on a selection criterion, he needs to use another selection program.
The speciﬁc objectives of this study were: (1) when selection is on a single trait,
estimate the genetic response of the trait undergoing selection, the correlated responses in
other single traits and in indexed merit ﬁrnction ; (2) when selection is on an indexed merit
value, estimate the expected response in that indexed merit, the expected correlated response
in each of its component traits, in traits which were not included in the merit index, and as in
other indexed merit ﬁrnction. The goal of this study is to provide information about all
genetic responses of economic single-traits or indexed merits to dairy producers in Taiwan
when they select semen from US. proven bulls with different selection criteria to improve
their herds. Moreover, using this information to set up breeding schemes will not only
improve in milk yield but also increase the total indexed merit to increase proﬁtability and

efﬁciency of a cow.

53

MATERIALS AND METHODS
Traits and indexed merit function
Dairy bulls in US. are currently being evaluated for a number of traits and indexed
merits. Single traits and their selection criteria included: (1) PTA for Milk, Fat, and
Protein; (2) Conformation - PTA for Type (ﬁnal score) and STA for functional type traits;
(3) PTA for SCS; (4) PTA for PL. PTA value is an estimate of genetics superiority or
inferiority that an animal will transmit to offspring.
Three yield merits are derived based on the current milk pricing system in the US:
MF$ = $.10] 70(PTA M)+$.58(PTA F)
= 5(PTA W625)+1(PTAF/22)
MFP$ = $0.031 (PTA M)+$0.8 (PTA F )+$ 2.00 (PTA P)
= 2.67(PTA M/625)+1(PTAF/22)+1.7(PTA P/19)
CY$ = -$.00220 (PTA M)+$1.98(PTA F )+$1 .716(PTA P)
= -(PTA M/625)+32(PTAF/22)+24(PTA p/19)
Two Total Merit selection indexes, which are currently in use, include:
NM$ = .7(MFP$)+$11.30 (PTA PL)-$28.22 (PTA SCS-3.2)
= 10 (PTA MF P$/71)+4(PTA PL/l 3.2)-l (PTA SCC/1.9)
(USDA-DHIA,1995).
TPI = [$7.89(PTA P)+$2.22(PTA F)+$71.43(PTA T) +$lZ.l8(UD)+
$6.5(F U) +$6.5(TP)+$6.5(RUH)+$4.87(RUW)+$4.06(UC)
= [3(PTAP/19.0)+1(PTAF/22.5)+1(PTAT/. 7)+. 65(UDC/. 8)

+. 35(FLC/. 85)] 50+576

54

Where UDC=. 3(UD)+. 16(FU)+. 16(TP)+. 16 (RUH)+. 12(RUW)+. 1(UC)
(U .S. Holstein Association. 1997)

Economic values assigned to PTA's in MF$ and in MFP$ were based on a milk price of
$12.20 per hundredweight of milk with 3.5% fat and 3.2% protein and differentials of 5.8
cents for fat and 14.7 cents for protein. This was the US average milk price for 1994 minus
the average hauling, assessments, and promotion charges (Norman et al.,l995). These
ﬁgures will change every year due to changes in milk prices. The net merit index (NM) was
based on MFP$ deducted for feed cost and on PL and SCS evaluation.

The second equation of each criterion was expressed as the relative weight for each
component trait, which was standardized by their genetic standard deviation. The
denominator terms were the standard deviations of PTAs for traits. The coefﬁcients in front
of PTAs were the relative weights of selection emphasis (Norman et al., 1995).

Calculation of genetic responses and correlated responses

The following formula in matrix notations was used to calculate the expected genetic
responses and correlated genetic responses in varied combination of all situation.
AG=(m'Cb)/(b'Pb)°‘5 * i / L
where AG : the expected genetic responses and correlated genetic responses in all single

traits and indexed merits

m: a vector of arbitrary weights for the traits in true value of a total indexed
merit ( net economic values for individual traits).

b : vector of a selection index weights for a given indexed traits in indexed
merit criterion for selection which will
maximize rn and true value of a trait or a merit.

P : vector of phenotypic value of individual traits expressed as deviations from
pertinent smallest subclass means.

C: (Co)variance matrix between p in indexed merit criterion for selection and g

in true value of a trait or a merit.
P: phenotypic (co)variance matrix.

55

L : generation interval
i : selection intensity

Genetic Parameters used in calculating genetic response
When calculating genetic response, some of genetic parameters are needed. In tis study,
genetic estimates are used from the literature which are the results of studies for US Holstein
population. Table 2-1 shows the estimates of heritability, phenotypic and genetic variances
to calculate expected genetic responses and correlated genetic responses. Genetic estimates
of 15 type traits were ﬁom Misztal et a]. (1992). Heritabilities for milk, fat, and protein were
also included in this paper which were higher than most of the other estimates of heritability
using the Sire Model (SM) (Misztal, 1992; Van Vleck et al.,1987; Visscher and Tompson,
1990). The reason of higher heritability was that SM accounts for the male genetic
variation whereas AM took both male and female genetic variation into account. If
selection intensity for males were greater than for females, then the male genetic variance
would be smaller. Another reason that higher heritabilities of yield traits were obtained in
Misztal study was the use of only registered animals. Heritability and genetic variance of
milk used in this study were .31 and 554,143 kg, respectively from Short and Lawlor
(1992)
The phenotypic covariance and genetic covariance between traits from Tables 2-1,2,
and 3 where used to construct of the phenotypic (co)variance matrix P and the genetic C
matrix (co)variance. Selection intensity was ﬁxed to be 1.4 (i.e. equivalent to selection the
top 20% of bulls) and selection accuracy was .8 for all traits as well as the generation
interval 6 years. The genetic responses and correlated responses are the results after on

round selection of generation.

56

Table 2-1. Literature estimates of heritability and variances for the three yield traits
(kg, 15 type traits (points), Somatic Cell Score (scores) and productive life (mo)

 

 

 

 

Traits Heritability Phenotypic Genetic
variance variance
Milk .312 1,787,5692 554,143?
Fat .30 3,333 1,000
Protein .30 2,000 600
Type ﬁnal score .29 13.4 3.8
Stature .42 58.2 24.5
Strength .29 46.8 13.6
Body Depth .35 48.1 17.
Dairy form .28 47.9 13.5
Rump Angular .28 24.0 6.8
Thirl width .26 44.5 11.7
Rear leg .16 40.1 6.2
Foot angle .13 34.6 4.5
Fore udder .24 45.7 10.8
Rear udder high .16 46.7 7.3
Rear udder width .19 45.3 8.6
Udder cleﬁ .10 27.9 2.8
Udder depth .25 16.3 4.1
Teat placement .22 33.2 7.4
Somatic Cell Score .er 1.25I .15l
Productive Life .0853 174.243 14.813

 

Without superscript from Misztal et al.1992. J. Dairy Sci.75: 544
1 from Schutz.1994. J. Dairy Sci. 77: 2113.

2 from Short and Lawlor.1992. J. Dairy Sci.75: 1987.

3 from VanRanden and Wiggans. 1995. J. Dairy Sci 78: 631.

57

Table 2-2. Literature estimates of genetic correlations between traits

 

 

 

 

Trait Milk Fat Protein Type SCS PL
Milk .69 .90 .16 .122 .443
Fat .69 .78 .33 .022 .235
Protein .90 .78 .27 .172 .255
Type ﬁnal score .16 .33 .27 -.304 .393
Stature .06 .13 .13 .75 -.ll" .063
Strength .02 .13 .1 .62 -06“ -.113
Body depth .15 .26 .23 .7 -.05‘ -063
Dairy form .59 .68 .67 .29 .184 .403
Rump angular .18 .01 .11 -. 15 -.08" .083
Thirl width .11 .12 .11 .65 -21“ -.023
Rear leg .09 -.01 .05 -.11 -.10‘ .003
Foot angle .10 .13 .17 .28 -.O6" .083
Fore udder -.31 -. 12 -.21 .54 -.414 .293
Rear udder height .19 .28 .32 .59 -.l9‘ .323
Rear udder width .31 .33 .4 .60 -.15‘ .313
Udder cleft .01 .17 .15 .52 -124 .303
Udder depth -.44 -.29 -.38 .33 -.424 .243
Teat placement -.03 .01 -.01 .56 -.314 .243
scs .122 .022 .172 -.30‘ -.10“
Productive Life .443 .235 .255 .393 -.104

 

Without superscript from Misztal et al. 1992. J. Dairy Sci.75: 544
2 from Schutz. 1994. J. Dairy SCi. 77:2113.

3 from Short and Lawlor. 1992. J. Dairy Sci.75:]987.

4 from Rogers et a1. 1991. J. Dairy Sci.74:1087

5 from Welper and Freeman. 1992. J. Dairy Sci:1342

58

Table 2-3. Literature estimates of phenotypic correlations between traits

 

 

 

 

Trait Milk Fat Protein Type SCS PL
Milk .75 .93 .21 -.1 2 .193
Fat .75 .81 .2 -072 .235
Protein .93 .81 .22 -012 .255
Type .21 .2 .22 -.O6" .193
Stature .09 .1 .11 .32 .004 .043
Strength .05 .07 .09 .24 .02“ .023
Body depth .13 .15 .16 .3 .024 .033
Dairy form .34 .29 .32 .2 .014 .093
Rump angular .04 .02 .03 -.04 .014 -.013
Thirl width .08 .01 .1 .24 .0 .023
Rear leg .01 .02 .01 -.02 .014 -.013
Foot angle .03 .03 .04 .06 -.01‘ .043
Fore udder -.05 -.02 -.05 .15 -.O8" .103
Rear udder height .17 .15 .18 .23 .03“ .113
Rear udder width .23 .20 25 .3 -.02‘ .113
Udder cleft .05 .05 .06 .12 -.O6" .103
Udder depth -.22 -.18 -.21 -.02 .09“ .063
Teat placement .04 .04 .03 .18 -.054 .093
scs -.102 -022 -.1 2 -.06‘ .064
Productive Life .193 .235 .255 .195 . 06“

 

Without superscript from Misztal et a]. 1992. J. Dairy Sci.75: 544
2 from Schutz. 1994. J. Dairy SCi. 77:2113.

3 from Short and Lawlor. 1992. J. Dairy Sci.75:]987.

4 from Rogers et al. 1991. J. Dairy Sci.74:1087

5 from Welper and Freeman. 1992. J. Dairy Sci:1342

RESULTS AND DISCUSSIONS

Genetic and correlated responses in single-trait to all selection criteria

Table 2-4 showed the genetic response and correlated response in three yield traits as
well as Somatic Cell Score (SCS) and productive Life (PL) to various selection criteria for a
constant selection intensity (20%) and generation interval (6 years) after one generation.
Selection on PTA M resulted in the largest annual increase for milk yield. But only slightly

lower responses were obtained from selection on MF$, MFP$, and NM$. The less

59

correlated responses in milk yield were obtained from selection on PTA F, PTA P, CY$, and
TPI. The smallest annual increase in milk yield was on PTA T and there was a negative
response in milk yield for PTA SCS. MF$ and MFP$ were designed for these producers
when the milk price was determined by milk with a fat differential and with both fat and
protein differential.

Table 2-4. Genetic responses in single traits to selection from single-trait and multiple-
trait indices

 

 

 

 

 

Selection Expected genetic response
Criteria Mllk Fat Protein Type SCS PL
(kg) 0(8) 0(8) (point) (score) (month)

PTA M 1 16.05 3.40 3.43 .05 .007 .264
PTA F 78.77 4.85 2.93 .10 .001 .136
PTAP 102.74 3.78 3.75 .08 .010 .148
PTA T 17.96 1.57 1.00 .29 -.O20 .231
PTA SCS -8.66 -0.06 -0.40 .05 -.037 .037
PTA PL 26.73 0.59 0.49 .06 -.003 .314
MF$ 115.37 3.54 3.45 .05 .007 .259
MFP$ 110.62 3.86 3.62 .07 .007 .213
CY$ 86.99 4.67 3.93 .10 .005 .118
NM 113.04 3.14 2.86 .05 .001 .383
TPI 65.38 3.60 2.73 .24 -.004 .211

 

Annual increase for fat yield would be the most from selection on PTA F. Most selection
criteria except PTA T, PTA SCS and PTA PL could lead to reasonable increase in fat yield
genetic response. Negative response in fat yield was also found selecting on PTA SCS. The
correlated response in fat yield to CY$ would be higher compared with other selection
criteria. The relative weight of selection emphasis in CY$ was 4 : 3 for fat yield and protein
yield. The smallest annual increase in fat yield was selecting on PTA PL and negative
response in fat yield was selecting on PTA SCS.

Genetic response in protein yield was obtained when CY$ was the selection criterion.

The correlated genetic responses in protein to PTA P, MFP$ and MF$ were also favorable.

60

The smallest annual increase in protein yield was on PTA PL and the negative response in
milk yield was on PTA SCS.

All selection criteria would increase SCS with much less from selection on NM$. This
indicated selection criteria which would produce higher genetic responses in yield traits also
lead SCS increase (.007). Selection on PTA P would increase SCS the most (.01). Genetic
improvement for reduced mastitis incidence is possible by selection for fewer somatic cells
count in milk from selection on PTA T and TPI. Because of the favorable correlation
between udder conformation and SCS, selection on TPI, PTA T and PTA SCS could
improve SCS genetically down and selection on NM$ would slow the rate of genetic
increase in SCS.

Reasonable genetic response for PL could be achieved by all selection criteria. Selection
on NM$ would obtain .38 month improvement per year and PTA PL and TPI would obtain
.20-.26 month genetic improvement per year. High genetic correlation (.44) between rrrilk
yield and PL would result in this improvement. The results indicated that future cows will be
healthier, higher producing, and more functional. Current average total herd life in US. was
65.2 mo for registered cows and 60.9 mo for grade cows (Short and Lawlor. 1992).

All selection criteria would lead to favorable responses in linear type traits with a few
exceptions including rear leg set, rump angle, fore udder attachment, udder depth and teat
placement.

Table 2-5 gave correlated genetic responses in linear body frame traits by
different selection criteria. All selection criteria would result in taller stature, more
strength, deeper body depth and steeper foot angle. Yet, with increased milk yield from

selection, it also would result in rear leg set sickle and rump angle more slope. Rump

6]

angle and rear leg side view traits in dairy cattle had been regarded as important traits. A
high pin rump would affect reproductive performance because of difﬁcult drainage
through the reproductive tract and possible cause calving problems. Rear leg set relates to
the durability of the legs and feet. Posty-legged cows would have too much stress on their
legs, caused by an aggravation of joints. Sickle-hocked cows would have too much stress
on the leg's muscles and tendons, both of which related to cow's longevity (Linear
Classiﬁcation Program, HF A. 1996). Short and Lawlor (1992) reported that rear leg side
view was not genetically correlated with longevity and indicated that cows with a certain
degree of sickliness in their rear legs stayed longer in the herd. Burke and Funk (1993)
showed that intermediate curvature of rear leg was associated with longer herd life.
Selection on TPI would lead to the rear leg slightly straighter (-.01) and PTA T would
lead much straighter legs (-.05). Selection on TPI would lead slightly higher pin rump (-
.01) and PTA T would lead to higher pin rump (-.04).

Table 2-5. Correlated genetic responses in body frame by different selection criteria

 

 

 

 

 

 

Selectction Expected genetic response ( score)

Criteria STA STR BD DF RA TW RL FA
PTA M .05 .01 .10 .34 .07 .06 .04 .03
PTA F .10 .07 .16 .38 .00 .06 -.00 .04
PTA P .10 .06 .14 .38 .04 .06 .02 .06
PTA SCS .05 .02 .02 -.O6 -.02 .06 .02 .01
PTA PL .02 -.03 -.02 ' .11 .01 -.01 .00 .13
MF$ .05 .02 . 10 .34 .06 .05 .03 .03
MFP$ .07 .04 .13 .37 .05 .06 .03 .04
CY$ .11 .08 .17 .40 .01 .06 .00 .05
NM .01 .01 .06 .27 .07 .07 .03 .02
PTA T .55 .34 .43 .46 -.05 .33 -.04 .09
TPI .27 .20 .30 .30 -.01 .18 -.01 .08

 

STA: Stature (- short + tall) TW : Thirl Width (- narrow + wide)
STR : Strength (- frail + strong) RL :Real leg (- posty + sicked)
BD : Body Depth (- shallow + deep) FA : Foot Angle (- low + steep)

DF : Dairy Form (- tight + open) RA : Rump Angle (- higher pin + slope)

62

Table 2-6. Correlated genetic responses in udder traits by different selection criteria

 

 

 

 

 

 

 

Selection Expected genetic response (score)
Criteria FUA RUH RUW UC UD TP
PTA M -.16 .08 .14 .00 -. 14 -.01
PTAF -.06 .12 .14 .04 -.08 .00
PTAP -.10 .13 .18 .04 -.12 -.01
PTA SCS .13 .05 .04 .02 .08 .03
PTA PL .07 .07 .07 .04 .03 .05
MF$ -.15 .08 .14 .01 -. 14 -.01
MFP$ -.13 .10 .16 .02 -.13 -.Ol
CY$ -.09 .14 .17 .05 -.10 .00
NM$ -.16 .04 .10 -.O2 -. 13 -.01
PTA T .26 .24 .26 .13 .10 .23
TPI .10 .22 .24 .09 .09 .14

FUA : Fore Udder Attachment (- loose + strong)

RUH : Rear Udder Height (- low + high)

RUW : Rear Udder Width (- narrow + wide)

UC :Udder Cleft (- weak + strong)

UD : Udder Depth (- below hooks +above hooks)

TP : Teat Placement (- outside + inside)

Table 2-6 shows the correlated genetic responses in udder traits by different selection
criteria. Selection on yield traits, yield merits and NM$ would result in slightly larger
udder, a loose fore udder attachment, higher and wider rear udder height and width, strong
udder cleft and udder depth gradually below hooks. The deeper udder and more loosely
attached fore udders are undesirable responses. Loosely attached fore udders usually
determined how the fore udder would be carried, increased the likelihood for injury, and
caused high mastitis incidence. While a degree of udder depth was necessary for capacity, a
deep udder was susceptible to injury and mastitis. (Linear classiﬁcation program, HFA.
1996). Selection on TPI would help to reduce or to eliminate these undesirable correlated
responses.

Selection on PTA PL would also lead to favorable response in all linear type traits.

63

Selection on PTA PL was not a "better" selection criterion to improve type traits, because
PTA PL was calculated by two components, one was direct trait using DHIA culling data
and another was indirect trait using linear type data to predict PL.

When selection was on TPI, response in milk yield was reduced about 54%. Misztal
(1992) used a restricted index to calculate maximum response in milk while maintaining
udder depth at its current value. Using this restricted index would result in 15% decrease in
genetic gain for milk yield Standardized weights for milk yield and udder depth were 70:30,
or approximately a 2:1 ratio. Rogers and McDaniel (1989) applied selection index
methods by utilizing involuntary culling in an aggregate genotype to represent the
composite costs of health, reproduction, and husbandry attributes. Traits considered in the
index were milk yield, udder depth, teat placement, and foot angle. Simultaneous
responses on increased milk yield, higher udder, closer teats, and steeper hooves were
reported. Rogers (1993) included SCS information in the selection index. Index
coefﬁcients were always negative for SCS and was positive for udder depth, teat
placement, and foot angle. Response on lower SCS, higher udders, and closer teat
placement were reported. Standardized index indicated that milk yield should be
emphasized three to four times as much as SCS, udder depth, teat placement, and foot
angle combined. Udder depth and SCS was the most useful non-yield traits.

Sale of milk was the major source of income for most dairy producers. Milk yield,
therefore, should receive heavy emphasis in selection programs. Selection on single-trait
selection criteria such as PTA T, PTA SCS, and PTA PL might result in low responses in
milk yield and should be not used as independent selection criteria. Due to their low

heritabilities, they could be included in multi-trait selection criterion.

64

Total genetic responses in the indexed merits

The greatest retum in any indexed merit was obtained by selection directly on that index.
Selection on PTA of production traits and other merit indices would lead to almost as much
response. Selection on PTA of SCS would result in very small positive responses in
indexed merits.

Table 2-7. gave total genetic responses in the indexed merits to different selection
criteria. The most improvement in an indexed merit would be from selection on its indexed
value. The annual expected response achieved in indexed merit of MP and MFP to selection
on MF$ were $13.78 and $13.61 per cow per year. But much higher responses were also
obtained from selection on PTA M, and NM$. The less correlated responses in indexed merit
of MP and MPF were obtained from selection on PTA F, PTA P, CY$, TPI.

Table 2-7. Correlated genetic responses in the indexed merits by different selection

 

 

 

 

 

criteria

Selection Expected genetic response (8)

Criteria MF MFP CY NM TPI
PTA M 13.77 13.37 12.37 12.10 35.67
PTA F 10.82 11.42 13.71 10.70 41.23
PTA P 12.64 13.33 14.45 10.70 43.40
PTA SCS - .91 -1.10 -.79 .70 4.26
PTA PL 3.06 2.54 1.97 5.38 13.27
MF$ 13.78 13.44 12.68 12.11 36.54
MFP$ 13.48 13.61 12.88 12.60 40.67
CY$ 11.56 12.45 14.48 11.42 44.89

.104 ...................... 1. 2.3.2 ............. 1.2.2.2. ............ 1.0.9.9.-------_-1239--_-_----_.2.9.28.--

PTAT 2 71 3.32 4.73 5 31 43 94
TPI 8 82 9 80 11.81 9 25 58 56

 

The annual expected response achieved in indexed merit of CY to selection on CY$ and
PTA P were $14.88 and $14.45 per cow per year. But much higher responses were also

obtained from selection on PTA F, PTA M, MFP$ and MF$. The less correlated responses in

65

indexed merit of CY were obtained ﬁ'om selection on TPI and NM$.

For indexed merit of NM, the most improvement in an indexed merit would be from
selection on NM$ ($12.80). Much higher responses were also obtained from selection on
MFP$, MF$ and PTA M. The less correlated responses were obtained from selection on
CY$ PTA F, PTA P, and TPI.

For indexed merit of TPI, the most improvement in an indexed merit would be from
selection on TPI ($58.56). Much higher responses were also obtained from selection on
CY$, PTA T PTA P, and PTA F. The less correlated responses were obtained from selection
on MF$ and PTA M.

Selection on PTA PL, PTA SCS and PTA T would lead small response in all indexed
merits. Yet, selection on PTA T would lead much higher response in indexed merit of

TPI ($43.94).

CONCLUSIONS AND IMPLICATIONS

The genetic response in milk yield would stay positive steadily, when any of the
selection criteria from USDA and Holstein Association USA was used, as long as milk
yield is included in the selection index. When milk was paid for its volume with a fat
differential, or with both a fat and protein differential, selection on PTA M could result in
largest response in milk yield with small difference on MF$, MFP$, and NM$. When the
payment for milk was based on fat or protein yield, selection on PTA F and CY$ could
get largest responses with small difference on MF$, MFP$, and NM$.

The study on correlated responses showed undesirable responses in SCS and in some

linear type traits, when selection emphasis was on increasing milk yield. These

66

undesirable responses may be caused by the health and fertility stress problems due to the
increased level of milk yield. If the severity of these undesirable responses increased, the
value of increased milk yields would be offset by the additional costs of handling for
these undesirable responses.

The selection criteria emphasizing yield traits generally would also increase SCS in
milk but to a much less extent from selection on NM$. Linear type traits such as rear leg
of side view, rump angle, fore udder attachment, and udder depth would result in
undesirable responses. The main breeding goal of dairy producer is to breed proﬁtable
cows, which will stay in the herd for several trouble-free lactations. If farmers' breeding
goal was to breed a more proﬁtable herd without concern about conformation, they might
use independent culling levels to select AI sires with high value of these selection
criteria. They should then scrutinized heavily for PTA SCS and for linear type traits such
as rear leg of side view, rump angle, fore udder attachment, udder depth and teat
placement. If farmers wanted to change cow conformation to meet their personal
standards, selection on TPI or additional selection pressure on linear type traits after

screening on yield selection criteria would be a better sire selection practice.

67

CHAPTER 3

PRACTICAL BREEDING SCHEMES TO MAXIMIZE GENETIC
IMPROVEMENT IN A DAIRY CATTLE POPULATION OF LIMITED SIZE

ABSTRACT

A dairy cattle population of limited size merits special biological and economical
considerations in designing an optimum and yet practical breeding strategy to maximize
genetic gain in milk-producing ability. In this study, the geographically isolated dairy
cattle population of 100,000 cows was the target population for genetic improvement, but
economical factors were not considered. Various breeding schemes were designed
considering practical conditions in the target population. All designed schemes involved
imported germplasms from the US. Schemes using local genetic resources included the
use of bulls from progeny test for AI services (PT/AI), the use of bulls with high pedigree
merit but untested for AI and natural services (untLB/AI/N S), and the use of a MOET
nucleus population to produce AI and NS bulls (MOET/AI/N S). Schemes using imported
germplasms included the use of embryos to produce untested AI bulls (untFEB), or to
produce AI bulls by paternal half-sib performance (STFEB) as well as the use of
imported semen from the top 40% (FS40) or top 20% (FS20) U.S. proven bulls to breed
local cows. Deterministic models were used to estimate the genetic level of milk
producing ability in each of the consecutive 25 years when a particular breeding scheme
was applied. The current genetic difference between the US. and the target population
was considered, and the effects of genotype by environment interaction were studied.

According to cumulated genetic progress after 25 year, the breeding schemes were

68

 

ranked in descending order as STF EB, untFEB, MOET/AI/N S, F820, F S40,
untLB/AI/N S, and PT/AI. All breeding schemes surpassed the genetic gain by the current
breeding scheme (CBS), except PT/AI, by 35%, 29%, 23%, 17%, 12%, and 5%,
respectively. Amongst the schemes utilizing local genetic resources, MOET/AI/N S was
better than CBS, PT/A] and LB/AI/N S by 19%, 40% and 13%, respectively. The schemes
using imported semen exclusively were competitive with the genetic progress after 25
years being 12% (FS40) and 17% (F820) higher than that resulted from CBS. The
ranking of schemes was not affected, although the magnitude of genetic gains were, by
adding a supplement of importation of 5% pregnant heifers, 5% yearling heifers, or
foreign semen (F 820 or FS40) to breed 30% or 50% cows. The ranking of schemes was
not inﬂuenced by genotype by environment interaction either. The use of imported
embryos to produce AI bulls by paternal half-sib performance (STFEB) with a
supplement of 5% imported pregnant heifers was the best scheme gaining 52% more than

CBS did.

INTRODUCTION
A breeding scheme is a practice of genetic selection with application of certain
reproductive technology or biotechnology. The goal of selection is to allow animals with
high breeding values to be parents of the next generation. The selection intensity,
accuracy of the parents' genetic evaluation, generation interval, and genetic variation will
determine the genetic progress (Rendel and Robertson, 1950). The shorter the generation
interval and the higher the selection intensity and selection accuracy, the more rapid

would be the genetic progress. Yet, in practice, it is difficult to do all these things at once

69

because a favorable change in one factor often made an unfavorable change in another.
For example, a decrease in generation interval causes a decrease in accuracy of selection.
To design breeding scheme is to balance the relationships among these factors so that
genetic improvement is optimized.

The ﬁrst and most important reproductive technology that had a major applied to
breeding scheme in dairy cattle was Artiﬁcial Insemination (A1) with frozen semen. The
introduction of A] offered excellent opportunities for the use of highly selected proven
bulls. Artiﬁcial Insemination and Progeny Testing (AIPT) has been the backbone of dairy
cattle improvement in developed dairy countries. This scheme identiﬁed and selected
bulls with superior genetic merit for milk, fat, protein and type through their relatives and
many daughters' performance. AIPT scheme could result in rates of genetic gain of 2% to
3% per year (Van Vleck, 1986). Yet, long generation intervals of progeny tested sires
have limited genetic progress in AIPT scheme.

Embryo transfer (ET) technology enables females to increase family sizes and
shortens the generation intervals. The use of Multiple Ovulation and Embryo Transfer
(MOET) as a potential tool for genetic improvement of dairy cattle in a nucleus herd was
ﬁrst outlined by Nicholas and Smith (1983). In a MOET nucleus scheme, selection on
young bull-dams to produce bulls could reduce generation interval and increase genetic
gain by 4% to 5 %, which would be superior to those achieved in conventional breeding
schemes (Dekkers and Shook, 1990).

Other reproductive technologies have been applied to breeding schemes, such as large
factorial mating breeding designs with use of in vitro embryo production (IVEP). The use

of physiological indicator traits and genetic marker as aids to select animals which was

70

called Marker Assisted Selection (MAS) was also recently studied by several geneticists.

To bring genetic material of higher genetic merit into a population is another way to
make genetic improvement, especially when the foreign population is genetically
superior to the local population. Introducing external germplasm include importing in the
forrrr of live animals, frozen semen, and frozen embryos.

There are many dairy cattle populations of limited size in the world that operate
independently either geographically or politically. Each, however, has unique features
such as management system, feeds and feeding, and climatic conditions that would
inﬂuence the level of milk production. Each also has made efforts to improve milk
production. Target is an island located in subtropical and tropical area. A dairy industry
had been developed since 1960. The demand for milk products has been increasing since
1980. However, the milk production has not met the market demand and large volume of
dairy products has been imported. Milk production on the island needs to increase, since
a large number of consumers prefer to have fresh milk produced locally. The dairy
industry in Target also has been spending a great deal of funds and efforts in the
importation of genetic material from abroad to make genetic progress. Govemmenta]
agencies, large private industries and small farmers were all involved in the massive
importation effort. Today, Target had a fairly well developed breeding and recording
infrastructure in place; i.e. technologies such as A], embryo transfer and DHI. At present
time, the dairy cow population is approximately 100,000 in size and will likely stay at
same considering the environment and resources for dairy farming. There does not exist a
long-term breeding policy for the dairy population in Target, which would accelerate and

sustain genetic progress in milk yield efficiently both biologically and economically. The

71

overall goal of this work was to seek an optimum breeding strategy tailor-made for
conditions and resources in Target. The speciﬁc objectives of this study were to design
alternative breeding strategies all practical under Target conditions and to use
deterministic models to calculate the expected genetic change for each of the alternative

breeding schemes in order to rank these proposed schemes.

MATERIALS AND METHODS

Descriptions and assumptions of the target population

Special characteristics of target population are described in chapter 1. For the
breeding study, some descriptions and assumptions are listed in Table 3-1. Long calving
interval (15 mo.) and high rate of involuntary cull result in only 65,000 milking cows
each year. Approximately 50% of cows are on the milk recording program (DHI). Over
80% of cows were bred by AI and the rest were mated to natural service (NS) bulls.
These NS bulls were of unknown genetic quality and it can be assumed that they do not
contribute to genetic progress. The age structure of cow population was derived from
current Target DHI database (Chang, 1997). Culling of cows was assumed to have no
effect on the rate of genetic progress and results in no selection for dams to cows path.
The average 305-2X-ME of milk yield was 6500kg and phenotypic SD was 800 kg
respectively. The heritability of milk yield was assumed to be .3.
Assumptions for the imported germplasm

All proposed breeding schemes involve imported germplasm. More than 80 % of
imported germplasm has been from Canada and the US. and the US. cattle and semen

are generally favored. Therefore, for simplicity, in this study, the US. was assumed to be

72

the sole source of imported genetic material. The U.S. dairy cattle population was

assumed genetically superior to the target population.

Table 3-1. Descriptions and assumptions of the dairy cattle population

 

Size cow population 100,000
Size milking cow population 63,000
Recorded cow population 50,000
Average lactation milk yield (305-2X-ME)(kg) 6,500
Phenotypic SD in milk yield (kg) 800
Genetic SD in milk yield (kg) 438
Heritability of milk yield .3
AI services per conception 2.7
Age at ﬁrst calving (yr.) 2.2
Calving interval (yr.) 1.3
Parity Age at
Calving %
1St 2.2 32
2"d 3.5 25
3rd 4.8 20
4th 6.1 15
5th + 7.4 8

 

The genetic difference in milk production between populations has been studied and
reported by INTERBULL. Target is not a member of INTERBULL. Therefore, we
assumed that a genetic superiority of 623 kg for the U.S. dairy population over Target
dairy population. This was approximated by (8578-6500)*.3=623 where .3 was the
heritability of milk production, 8575 was the average of milk production of the U.S. dairy
population (Wiggans, 1997) and 6500 was the average of milk production of target dairy
population (Chang, 1997).

Appendix A Table 1 shows the AIPL - USDA report about the genetic trend for
Holstein (1998). From this report, the genetic mean for the U.S. cows was lagged behind
that for the U.S. progeny-selected bulls by 400 kg (880 lb) in 1990. So the genetic

superiority of 1023 kg for the US progeny tested bulls over target population was

73

approximated. Since 1989, USDA has used an animal model to evaluate the breeding
value for cows and bulls. Genetic evaluation is much more accurate than has ever been
done. From 1990 to 1996 in the U.S. dairy population, the average annual genetic change
was approximated to be 120kg for cows and 126kg for bulls. It was assumed that these
constant annual genetic gains in milk yield for cows and bulls would be the same for 25
years in this study. Imported germplasm will be in forms of semen, embryos and live
animals with different genetic levels. The notation for imported germplasm will be:

1) F S stands for the imported foreign semen selected from progeny-tested active
proven bulls list according to PTA M with REL over 70% in the U.S. FSS, for example,
stands for the imported semen selected from the top 5% proven bulls by PTA M.

2) FEB stands for bulls from imported foreign embryos.

3) FH stands for imports of foreign heifers. Fng stands for imports of foreign
pregnant heifers, and F ylH stands for imports of foreign yearling heifers.

Current Breeding Scheme (CBS)

Current Breeding Scheme summarized the real current breeding practices in the target
population.

(A). Local selected but untested bulls to mate to 20% of cows by A]:

A governmental bull station was established to raise young bulls for AI purpose. 50
second lactation cows selected from top 5% on PTA M were selected as bull-dam
candidates and were contracted to mate to FSlO with 81% REL. These mating would
produce 20 bull calves each year. Nine out of these 20 bull calves were purchased based

on their development and health at 6 months of age and raised at bull station.

74

Three of nine bulls were selected based on PA (Parent Average) of PTA M and were
raise to 1.5 years old. They would be used in service from 1.5 to 4.5 year of age for three
years. The culling rate of untested bulls was 33%. Then 6 untested bulls were kept in Al
services each year to mate to 20% of cows (2.7 doses *20,000 *1.05 =56,000 doses
semen for AI. Each bull produces 10,400 doses of semen per year. (200 doses/bull/week).
(B). Imported semen mated to 60% of cows by Al:

The 170,100 doses of F850 with 70% REL was imported per year to mate to 60% of
cows in the population (2.7 doses *60, 000 cows *1.05 where 2.7 was service times per
conception; 1.05 was for contingency to allow for losses from damage semen during
handling semen).

(C). Natural service bulls (N S) with unknown genetic merit settled 20% of cows:

Most of herds keep natural service bulls with unknown genetic merit to sire 20% of
cows. The intent of NS was to impose the reproductive efﬁciency. Though some of their
sires were foreign, their dams were high producing cows in the herd without records, they
were kept in the herds and used as NS bulls from 1.5 to 3.5 year of age for two years
before culling. So they were assumed to contribute nothing to genetic improvement to the
next generation.

Alternative breeding schemes

Considering practical situation in target population, three categories of alternative
breeding schemes were design. Since it was unlikely in practice that all cows would be
bred by AI, assumed that in each proposed breeding scheme, 20% of cows were mated to
unknown genetic quality bulls by natural service. The rest 80% of cows were mated to

the bulls produced from each of the proposed breeding schemes through A]. Natural

75

service (N S) bulls would not contribute genetic progress to the population except when
they were from contract matings from some of the alternative breeding schemes. In those
breeding schemes, contract mating were outstanding performance cows with complete
pedigree records mated to speciﬁc superior proven bulls. The young bulls from contract
mating with known estimated genetic merit would be used as NS, and would contribute
genetic progress to the population. These N8 bulls, however, would stay as NS bulls,
each herd with one NS and would only mate 30 cows a year. They could not be used
extensively and could not be collected semen ﬁ'om them due to government regulation.

For a dairy population, the number of bulls required for AI purpose per year was
given by (Lindhe, 1968):

PCP "' CP* SPC

 

No. of bulls for AI =
Straws collected per bull per year

Where PCP was the proportion of cow population mated to bulls by AI (80%). CP
was the cow population (100,000), and SPC was the number of services per conception
(2.7). In this study, if 20,800 doses of semen were collected from each progeny tested
bull per year (400 doses *52 weeks), the number of bulls required for AI purpose in
these dairy population was 10. For the schemes to produce untested elite bulls for target
genetic improvement the rate of usage per AI bulls was reduced to 10,400 per year and
the number of bulls required for AI purpose was 20.

The selection criterion for all breeding schemes is PTA M. It is estimate of genetic
superiority (inferiority) that an animal will transmit to offspring. The accuracy of PTA M
is reliability (REL). It is measure of amount of information in evaluation. For the young

animals, Parent Average (PA) is to use to identify superior young animals early. It is

76

calculated as one-half of his sire's PTA M plus one-half of his dam's PTA M.

The common feature of all the alternative breeding schemes in this study was that
sires of bulls (SB) in breeding schemes were selected from progeny tested proven bulls in
the U.S. Table 3-2(a) and 3-2(b) gave summaries of abbreviation and deﬁnition of
breeding schemes. Alternative breeding schemes therefore could be divided into three
categories:

1. Breeding schemes using local bull-dams mated to foreign semen with high genetic
merit as bulls sires to produce selected A] or NS bulls. The related schemes were PT/AI,
LB/AI/NS and MOET/AI/NS.

2. Breeding schemes using germplasms to produce AI bulls or exclusively using semen
from progeny tested foreign proven bulls to mate cow population. These schemes were
untFEB, STFEB, F840, and F820.

3. Breeding schemes in category 1 and 2 with supplements of two kinds of imports. The
ﬁrst one imports were importation of foreign pregnant heifers or yearling heifers to
replace 5% of the cow population each year. The second imports were supplemented by
importation of F S with different genetic level to mate to breed 30% or 50% cows.
Breeding schemes using local bull-dams and foreign bull-sires to produce AI bulls:

(1) Progeny tested bulls for AI (PT/AI)

The F85 from ten foreign proven bulls with 90% REL would be used as sires of bulls.
Local proven bulls were not used as sires of sons due to smaller effective population size,
lower genetic level and lower accuracy than that of F85. Eighty cows of second lactation
in DHI herds with their genetic evaluation were in the top 5% would be selected as bull-

dams and contract to mate to F 85. 30 bull calves would be produced. 20 bull calves were

77

purchased based on their development and health condition and raised in the bull station.
At age of 1.5, they mated all 20% of AI cow population. Each young bull had at least 100
daughters in DHI herds. Each young bull had at least 100 daughters with one lactation
record and did not include any second crop daughters when they were proven. All records
of these bulls' daughters were connected when doing genetic evaluation. Four out of the
20 sampling young bulls based on the rank of their progeny ranking would be selected.

From 2 years of age, semen of each young bull were collected, frozen, and stored
regularly and routinely. Once 55,000 straw semen from each bull had been collected and
stored and the bull was slaughtered. When the genetic merit proof of each bull become
available, semen from low proof bulls were disposed. The stored frozen semen from
selected proven bulls was provided to target population for regular service. They would
mate to 60% of cows (20% of cows mated by testing young bulls). The generation
interval was 7 years for cow-sires.

(2) Untested Local elite bulls for AI and NS (untLB/AI INS):

Each year, 300 cows of second lactation in DHI herds with their genetic evaluation
were in the top 5% would be selected as bull-dams and contract to mate to F85. One
hundred and twenty bull calves would be ranked by the estimated Parent Average for
PTA M. Top 8 bull calves were purchased and raised in the bull station and the rest 100
bulls would be used as NS bulls. The rate of usage per A] bull was reduced to 10,400 per
year to avoid risk. Therefore 20 untested young bulls were required to sire of 80% cows
by AI. Semen collection starts at 15 months and bulls were used for four years from age
of 1.5 year to 4.5 years. Each N8 bull would only mate to 30 cows per year ﬁ'om age of

1.5 year to 3.5 year and therefore only 6 % of cows population would mate to NS bulls

78

per year (30 cows/bull*200 bulls = 6000 cows).

Table 3-2(a). Abbreviation and deﬁnition of breeding schemes using local bull-dams

 

 

Abbreviation Scheme description

CBS Current Breeding Scheme.
The real current breeding practices in the dairy cattle population in
Target which included:

20% of cows mated to local untested elite bulls by AI
60% of cows mated to imported semen(F850)
20% of cows mated to unknown natural service bulls

PT/AI Ofﬁcial progeny tested bulls for AI.
10 progeny tested bulls were from contract mating that the imported
semen of F85 mated to top 5 % PTA M cows. Bulls were proven by
their 100 daughters' performance. Semen from these bulls would be
collected and stored after sampling. They would mate 80% of target
cow population after proven.

untLB/AI/NS Untested Local elite bulls for AI and NS.
20 AI bulls and 200 NS bulls were produced from the same contract
mating as PT/AI. 20 AI bulls were selected by the rank of estimated
Parent Average (PA) for PTA M to sire of 80% cows by A1. 200 NS
bulls were in 200 herds. Each bull mated to 30 cows in the herd by
NS and would contribute to 6% of target pomrlation genetic progress.

MOET/ AI/NS Nucleus population with Multiple Ovulation and Embryo Transfer to
produce AI bulls and NS bulls.
20 AI bulls and 200 N8 bulls were produced from the controlled
Nucleus population bred by Multiple Ovulation and Embryo Transfer
(MOET) techniques. A] bulls were selected by the estimated Parent
Average (PA) for PTA M to sire of 80% cows by A1. 200 NS bulls
from MOET were used in 200 herds. Each bull mated to 30 cows in
the herd by NS and would contribute to 6% of target population
genetic progress. '

79

Table 3-2(b). Abbreviation and deﬁnition of alternative breeding schemes using
imported germplasm and their features.

 

Abbreviation

Scheme description

 

untFEB/AI

STFEB/AI

F840

F820

/AI+F 840-3 0%

/AI+F820-30%

Untested bulls from imported foreign embryos for AI

10 AI untested bulls were produced from imported frozen embryos
but transferred in local recipients. These embryos were from
planned mating of elite cows and bulls in the exporting country and
selected on the estimated Parent Average (PA) for PTA M. They
would mate 80 % of target cow population.

AI Bulls from imported embryos tested by paternal half-sibs

10 AI bulls were produced from imported frozen embryos but
transferred in local recipients. These embryos were from planed
mating of elite cows and bulls in the exporting country. 600 doses
semen of the sires of embryos would also imported at the same time
so that there would be 100 paternal half-sib to the bulls. Bulls were
selected by the half-sister performance in the local population.

A total of 227,000 doses frozen semen were imported each year to
mate 80% target cow population. The imported semen were selected
from the top 40% proven bulls list on PTA M with 70% REL from
the exporting countries to sire cows.

A total of 227,000 doses frozen semen were imported each year to
mate 80% target cow population. The imported semen were selected
from the top 20% proven bulls list on PTA M with 70% REL from
the exporting countries to sire cows.

Import of F 840 to bred 30% cow population to supplement AI.

F840 semen were imported to breed 30% of the target cow
population to supplement to above alternative breeding schemes.
Import of F820 to bred 30% cow population to supplement AI.

F820 semen were imported to breed 30% of the target cow

population to supplement to above alternative breeding schemes.

 

/AI+FyrH

Importation of foreign yearling heifers as 5% of the replacement
females supplemented to AI scheme. Selection for FyrH was based
on the Parent Average (PA) of PTA M being greater than that of
average in exporting cow population.

 

/AI+Fng

Importation of foreign pregnant heifers as 5% of the replacement
females supplemented to AI scheme. Selection for FyrH was based
on the Parent Average (PA) of PTA M being greater than that of
average in exporting cow population.

80

 

(3) Nucleus population with Multiple Ovulation and Embryo Transfer to produce AI bulls
and NS bulls (MOET/AI/NS):

Four nucleus herds would constitute controlled nucleus population with 300 cows and
100 heifers in size per herd. The feeding and management of these four herds would be
kept similar and controlled. They were on DHI with accurate and complete performance
and pedigree records. All management and feeding practices would be recorded in all
herds. The nucleus population was genetically better than the base population and would
be kept open to use foreign semen to mate the cows. In each nucleus herd, top 25 heifers
selected on PA of PTA M in the herd would be used as donors each year. Donors were
superovulated on the individual farms, mated to F85 with 90% REL, giving 4 embryos
per ﬂush. 400 embryos would be transferred to the recipients that were healthy heifers.
For one embryo to be transfer, three synchronized heifers were needed.

80 another 40 dairy farms would collaborate with the nucleus herds to provide 20
synchronized heifers for embryos transferring. With a pregnancy rate of 60 %, a calf
survival rate of 85 %, and a sex ratio of 50 %, 110 bull calves and 110 female calves
would be produced each year (4 embryos * .65 * .85 "' .5 * 100 donors = 110). Bull
calves would be ranked by the estimated Parent Average (PA) for PTA M.

Top 10 bulls from 10 families would be purchased and kept for AI and the rest 100
bulls would be sold as NS bulls.

In each nucleus herds, 130 female calves would be born each year in each nucleus
herd (15 from MOET heifers, 30 from heifers, and 100 from A1 cows 300*.75*.5=112).
Only 300 cows and 100 female calves would be kept in the herd according to their PTA

M and the rest would be sold or culled. The cows in the nucleus herds would be mated by

81

F810 with 80% REL to make genetic improvement.

An initial genetic superiority in milk yield of 300 kg for the nucleus population over
the target population was approximated by (7500-6500)*.3=300 where 7500 and 6500
was the average of milk production of the nucleus population and target population
respectively; .3 was the heritability of milk production.

Breeding schemes using imported germplasms to produce AI bulls:
(1) Untested bulls from imported foreign embryos for AI (untFEB/AI)

The value of estimated Parent Average for PTA M of these embryos was in the top
20% PTAM of AI proven bulls. The sires of embryos were with 90% REL from progeny
tested bulls and the dams were with 50% REL.

Each year, 4 young bulls would be produced every year from the imported frozen
embryos. Since the embryos would be frozen embryos, success rate was assumed to be
40%. 24 frozen embryos would be imported per year (24*0.5*0.4*1.05*0.85). The bulls
were raised at bull station and used as untested bulls to collect semen for AI services
from age 1.5 year to 4.5 year for four years. The culling rate of bulls per year was 25%.
Thus, 10 untested bulls were kept per year to collect semen to mate 80% of cows through
AI. Each bull would produce 20,800 doses of semen per year (400 doses/sire/week).

(2) Bulls from imported embryos selected for Al on patema] half-sib evaluations
(STFEB/AI)

Selection based on the half-sister performance had the disadvantage that the
correlation between the average breeding value of half-sibs and the breeding value of the
bulls was only half the correlation between the average breeding value of the progeny and

the bull's breeding value. Yet, the advantage of this selection was the age of bulls at

82

selection was reduced by 3 years. The genetic quality of imported embryos for this
scheme were the same as untFEB. More 600 doses of semen of embryo's sire were
imported as cow-sires, 100 half sibs to the young bulls were produced (100 half sibs *
2.7* 2*1.05 where 2 was the sex ratio). These sibs would have 1St lactation records
completed when they and their brothers were 3 year of age. The accuracy of evaluation
for bulls based on dam with one lactation record and 100 half-sibs. Then, selection
accuracy would be increase. Four out of the 10 young bulls based on paternal half sibs
test ranking would be selected. Under this situation, 56 frozen embryos would be
imported per year (56 * .5 *.8 * .4 * 1.05 * .85). The bulls were raised at bull station and
used as selected bulls to collect semen for AI services from age 3 year to 5 year for four
years. The culling rate of proven bulls was 25%. 10 bulls with different age were kept per
year to AI 80% of cows.

(3) Total reliance on foreign semen (F8)

A total of 227,000 doses frozen semen (80,000 Cows*2.7*l.05) were imported each
year to mate 80% target cow population. Two schemes of semen importation were
studied:

1) F 840 stands for the imported semen selected from the top 40% proven bulls list
on PTA M with 70% REL from the exporting countries to sire cows.
2) F820 stands for the imported semen selected from the top 20% proven bulls list
on PTA M with 70% REL from the exporting countries to sire cows.
Individual breeding schemes supplemented by additional imports :
Two kinds of additional imports supplemented to individual breeding schemes. One is

semen importation from top 20% or 40% to breed 30% or 50% of the target cow

83

population. The second is 5% heifers imported supplement to individuals.

1) Imported foreign semen to supplement AI: AI/+FS40-30% and AI/+FS40-50%
stand for importation of F 840 semen to breed 30% and 50% of the target cow population
to supplement above breeding schemes respectively. AI/+F820-30% and AI/+F820-50%
stand for importation of F20 semen to breed 30% and 50% of the target cow population
to supplement above breeding schemes respectively.

2) Imported foreign heifers to supplement A]: Female animals from foreign country
were imported to replace 5% of cows in target population. AI/+FyrH stands for
importation of foreign yearling heifers as 5% of the replacement females supplement AI
scheme. AI/+Fng5 stands for importation of foreign pregnant heifers as 5% of the
replacement females supplement AI schemes. AI/+FylH stands for importation of foreign
yearling heifers as 5% of the replacement females supplement AI schemes.

Imported heifers were selected based on the Parent Average (PA) of PTA M being
greater than that of average in exporting cow population. Yearling heifers were imported
at their age of one year. Pregnant heifers were mated to F850 before they were imported.
When they were imported, age range of them from 20 to 24 mo with 3 to 5 mo
pregnancy.

The rate of genetic change under a breeding scheme
1. Fundamentals of the rate of genetic change

Rendel and Robertson (1950) predicted genetic change based on four selection
pathways: Sires to bulls (SB), Sires to cows (SC), Dams to bulls (DB), and Dams to cows
(DC). The expected rate of genetic response as the result of parents selected was

calculated for each pathway separately,

84

G=i* r*O'a/L

Where AG : expected genetic response
i : selection intensity
r: accuracy of genetic evaluation
Caddy“... : genetic standard deviation for milk yield.
L: generation interval.

Table 3-3. The selection intensity in each selection path for alternative breeding schemes

 

 

 

Breeding Selection intensity
scheme
(abbreviation) SB DB SC DC
Control (C)- untAI bull(20%) 1.75 - .33 0
F 850(60%) - - .80 O
PT/AI 2.06 2.06 1.33 0
UntLB/AI/NS 2.06 2.06 1 .73 0
MOET/ AI/NS
Nucleus population 1.26 1.26 1.75 1.16
Target population - - 1.73 0
untFEB/AI 2.06 1 .26 0 0
STFEB/AI - - .6 0
F840 - - .83 0
F820 - - .83 0

 

One DE was the amount of information contributed to a parent by a standard
daughter that had one record and an inﬁnite number of management group mates. Total
DE for an Animal (DE Animal) was the sum of DE from PA (DEPA), Own yield (DE yield)
and progeny adjusted for mates (DE p.08.mm). Daughter Equivalent (DE) contributed to

REL using Animal Model (AM) procedures by different sources of information shown as

 

 

the following:
Information available Daughter
Equivalents
Parents Sire with70% REL and dam with 30% REL 4.7
Self 1 lactation record 4.7
Daughter 1 lactation record in one herd 1.0

 

85

DEanimal =DEPA + DEyleld +DEprog- mate]
RELanimal = DE animal/ (DEanimar+ 14)

REL used in this study would not include any information on second crop daughters.
For example, a bull had 50 lactating daughters. REL of his parent evaluation were 90%
for sire and 30% for dam. Then, REL for the bull was 79% ((4.7+50)/ (4.7+50+14) =.
79). If the progeny information (50 daughters) were not available, the REL would be 25%
(4.7/ (4.7+14)=0.25). The accuracy of selection would be 0.88 and 0.5 respectively for
the above example. Table 3-4 gave the accuracy of selection in each selection path for
breeding schemes.

Table 3-4. The accuracy of selection in each selection path for breeding schemes

 

 

 

Breeding Accuracy of selection
scheme
(abbreviation) SB DB SC DC
Control (C)- untAI (20%) .9 .63 .5 .35-.7
F850(60%) - - .83 .35-.7
PT/AI .95 .63 .93 .35-.7
UntLB/AI/NS .95 .63 .5 .35-.7
MOET/ AI/NS
Nucleus population 9 .5 .9 .35-.7
Target population - - .5 .35-.7
untFEB/AI .9 .63 .5 .35-.7
STFEB/AI 9 .63 .6 .35-.7
F840 - - .83 .35-.7
F820 - - .83 .35-.7

 

Prediction of genetic value of an animal with no records was usually from evaluations
of its sire and dam. The accuracy of evaluation was a simple function of the accuracy of

the evaluations of the two parents.

86

 

. = 1/2‘/(RELSM + RELD..)

For STFEB schemes, the accuracy of evaluation for bulls was calculated using
sibling and parents information. Because no Daughter Equivalents for sibling
information, the accuracy of evaluation for bulls was calculated using selection index
theory.

(c). Generation interval (L): L was the average age of the parents when their offspring
were born. The following were the calculation of L. Table 3-5 showed the generation
interval in each selection path for breeding schemes

1) SC: For the alternative breeding schemes involving bull selection, untested bulls
mate to cow population from 1.5 to 4.5 years and the generation intervals were 3 to 6
years. For the progeny tested bulls, his daughters would be born at his age of year 3. His
daughters ﬁnished their ﬁrst lactation at his age of year 6. For local proven bulls, they
sire to cows and his offspring born at his age of year 7 (Lsc =7). For imported semen as
sire of cows, they were imported at their age of year 7 and mated to cows and their
daughters were born at his age of year8 (Lsc =8).

2) DB: For the alternative breeding schemes involving bull-dam selection, a cow
must have at least one own record in her genetic evaluation to be selected as a bull-darn.
This evaluation would be obtained at the beginning of her second lactation. The contract
mating would take place during her second lactation. Her son would be born at her third
lactation calving and her age at that time was 5 years old (LDB =5).

3) SB: For the alternative breeding schemes involving bull-sire selection, imported

semen of foreign proven sire mated to bull-dams at their age of 7. Their sons born at their

87

 

age of 8 (L33 =8).
4) DC: The generation intervals for DC path would be 3 to 8 year at their lactation
calving for all breeding schemes.

Table 3-5. The generation interval in each selection path for breeding schemes

 

 

 

Breeding Generation interval (years)
scheme
(abbreviation) SB DB SC D C
Control (C)- untAI (20%) 9 5 3-5 3-8
F 850(60%) - - 8 3-8
PT/AI 9 5 7 3-8
UntLB/AI/NS 9 5 3-5 3-8
MOET/ AI/NS
Nucleus population 9 3 9 3-8
Target population - - 3-5 3-8
untFEB/AI 9 5 3-5 3-8
STFEB/AI 9 5 4-5 3-8
F840 - - 9 3-8
F820 - - 9 3~8

 

(d). Additive genetic standard deviation (0,)

Additive genetic standard deviation (0,) refers to the variability of breeding values
within a population for milk yield. It was somewhat difﬁcult to change. Bulmer (1971)
showed that the additive genetic variation was temporarily reduced by the selection of
parents and ancestors, but regenerates quickly when selection was relaxed.

2. Calculations of genetic progress

The annual genetic progress under each of the breeding schemes over 25 years was
evaluated using deterministic models. The generations were considered overlapping
because at any given time, cows were of different ages and the selection of parents was a
continuous process. Under this situation, animals of different age classes were regarded

as coming from different distributions of genetic merits with different means and

88

 

variances due to genetic trend.

The time-dependent population inventory approach presented in Ducrocq and Quaas
(1988) was used (Method DQ). Method DQ approach was suitable for predicting
response for a dynamic selection system. To calculate the genetic progress in the cow
population under a particular breeding scheme the ﬁrst step was to identify the proportion
of different age cow groups in the population at year I, each group with the same age. The
genetic progress of the cow population for each breeding scheme was weighted genetic
merit mean for cows with different ages. Weighting was on the proportion of the cows'
age group. The genetic merit mean for newborn animals could be obtained by the
averaging genetic merits of their parents. The average genetic merit of their parents was
the selection differential of their parents plus genetic mean of the animals at the birth year
of their parent.

The Method DQ also allowed the use of local or imported germplasm in each of the
selection paths. In calculating genetic merit for imported germplasm, the initial genetic
difference between the exporting and importing populations, the genetic trend in
exporting, and the different levels of genotype by environment interaction were
considered. The reduction of additive genetic variance due to selection (Bulmer effect)
and inbreeding depression was not considered, because all breeding schemes in this study
involved importation of germplasm from foreign country.

At year t, the genetic merit of the newborn animals was determined by the genetic
merit of the parents selected at year (t -L). The genetic merit of the parents selected at
year (t -L) was calculated for each pathway separately as following equation:

In (r-L) = H (t—L) + AG (t-L) [32]

89

 

Where m (14,) : the genetic merit of parents selected at year (t-L) and produced
offspring at year t.
L: the age of the parents when offspring was born at year t.
p. (g -L): the genetic merit mean of animals at year (t -L).
AG(.-L): selection response of the selected parents at year (t -L).
The genetic merit of an animal born at year t was the average genetic merit of the parent :
(m 8(t- Ls)+ m D(l- 1.6)) / 2 [3.3]
Where m 5((. Ls) and m D0 - Lo); the genetic merit of sire selected at year (t-Ls) and
the genetic merit of dam selected at year (t-LD) respectively.
m 80 - Ls) =llsu - Ls) + AGsu 4.5)
mou-LDFHDU -LD)+AGD(, -1...)
The average genetic level for newborn animals (bulls or cows) at year t :
Me (1) = (201 SCQ-Lsc) * m scq —Lsc )+ 20¢ DCQ-le) * m ch— Loo) / 2 [3.4]
M B (p = (20L sat; -Lsa) * m sea —Lsn)+ 20% DBQ —Lou) * m DBQ 4.126)) / 2
where u c (D and it B (03 the average genetic level of new born female animals and
male animals at year g,
m sea 4,.) and m 53(1_Lsp) : genetic merit of sire selected at year Q - Lsc ) and at year
(1 - L53) to be SC and SB.
Lsc , LDC ,LSB ,and LDB : the age of SC and DC as well as SB and DB at year 1
when their offspring were born.
use: the proportion of sires of cows with different age at year g.
orpcr the proportion of dams of cows with different age at year 1.
For the all schemes, selection and mating started year 0. Bulls were born and

selected at year 1. For the untested young bulls, they would mate to target cow population

at year 3 when they were 2 years of age. In PT/AI, before year7, assume all Al cow

90

 

population were bred by young sampling bulls which age were 2 to 3 years. From year 7,
ﬁrst group of proven bulls at their age of 7 was bred to 60% of AI cow population and the
young sampling bulls mate to the rest of 20% AI cows population. From year 4 on for
untested bulls and from year 8 on for progeny tested bulls, the bulls produced from each
breeding scheme mated 80% of target cow population.
3. Calculating the genetic progress of cow population at year t

Age (1) =20)(i, t) "‘ M c (t -0 [35]

where Agc(t) : the genetic progress of cow population at year t.
co(i, t): the proportion of the different cow age groups (i= 2,3,4,5,6,7) at year t.

The equation [3.5] meant the genetic progress of the cow population at year t for each
breeding scheme was weighted genetic levels for cow groups with different ages at year 1
compared with that at year 0 which was set up to 0.
Genotype by environment interaction between populations
When a breeding scheme was based on imports of genetic material from the foreign
country, genotype by environment interaction was considered. The product-moment
correlation between breeding value of sires estimated in the U.S. and in target population
could be used as a measure of genotype by environment interaction (Petersen, 1975;
Danell, 1982). Genetic progress should be adjusted for genotype by environment
interaction and it could be regarded as a correlated response (Falconer, 1989). The
correlated response for germplasms selected in the U.S. and used in Target would be:
AG Target = ius * 1‘05" 0a "target" rg us, Target
AG Target: selection response for germplasms selected in the
U.S.and responded in Target
ius: the selection intensity for the imported genetic material in

the US;
rug : the selection accuracy of genetic evaluation in the US;

91

(3'a Tamer: the additive genetic standard deviation for milk yield in
Target;
rg us, Tamer : the genetic correlation between the US and Target
Up to now, BLUP genetic evaluation system in Target had not been established.
Information on genotype by environment interaction (G X E) was not available. So the
value for the genetic correlation between the two countries was assumed to be .5 and .9
respectively to investigate the effect of r8 on the ranking of breeding schemes. The
genetic correlation between nucleus population and base population was assumed to be

1.0 for the MOET nucleus scheme. That rg assumed to be .9 meant there was little GXE

while rg were .7 and .5 meant there existed moderate to considerable GXE respectively.

RESULTSAND DISCUSSIONS

Estimated genetic trend for the exporting country

The initial diﬁerence in genetic merit mean between the target cow population and
the U.S. cows population and the US progeny tested bulls was assumed to be 623 kg and
1003 kg, respectively. The average annual genetic gain was assumed to be 120kg and
126kg for the U.S. cow population and the U.S. progeny tested proven bulls population,
respectively. Based on these assumptions, Table 3-6 showed the trend in the U.S. dairy
populations and genetic merit of gemplasm imported from the U8 into target population
which included F850, imported heifers and new born females from imported pregnant

heifers.

92

 

Table 3-6. Assumed genetic trend in the U.S. dairy populations and genetic merit of
gemplasm imported from the US into target population

 

Year The U.S. pop. Germplasm imported into target pop.

 

 

 

AI Cows F850 Imported New born females
bulls heifers from imported
pregnant heifers
-3 645 263 223 385 -
-2 771 383 349 505 -
-1 897 503 475 625 -
0 1023 623 601 745 427
5 1653 1223 1231 1345 1042
10 2283 1823 1861 1945 1657
15 2913 2423 2491 2545 2272
20 3543 3023 3121 3145 2887
25 4173 3623 3651 3740 3502

 

Estimated genetic levels and genetic progress

Table 3-7 and Table 3-8 showed the results at speciﬁc year for CBS and nucleus
population. The assumption of genetic merit mean in milk yield at year 0 was set up 0.
All selection and importation started at year 0. The progress of the genetic merit in milk
yield of new born males from contract mating in CBS, PT/AI, LB/AI/NS, MOET/AI/NS,
unt FEB, and STFEB started at year 1. These bulls started to mate to the cow population
at year 3. So the progress of genetic merit in milk yield of new born females started at
year 4 and there was no genetic response on the cow population until year 7 when these
female oﬁ‘spring calved and milking. The response was unstable in the earlier period until
the change of the number of cows and bulls with different ages in the population were
stable. Figure 3-1 gave the trends of estimated genetic progress of target cow population
over 25 years under proposed breeding schemes. The existence of the time lag between
the operation of selection and their effect on the cow population were nicely visualized.

The trends of genetic levels and genetic progress were non-linear over 25 years.

93

Table 3-7. Estimated annual genetic progress of breeding stock and the target cow

population for the Current Breeding Scheme (CBS)

 

Year Genetic merit mean of breeding stock at birth

 

Genetic progress

 

F850 untAI Bulls Females In target pop.
0 223 0 0 0
1 349 542 131 0
4 727 731 344 10
7 1105 1009 609 125
15 2113 1878 1449 702
20 2743 2464 2022 1 132
25 3373 3065 2613 1589

 

Table 3-8 Estimated genetic progress in the nucleus population for the MOET/AI/NS

 

Year Genetic merit mean of breeding stock at birth

 

Genetic progress

 

AI bulls (F820) Females mJCICuS population
0 0 300 300
l 625 400 300
4 1003 589 427
7 l 3 81 810 617
1 5 23 89 1766 155]
20 3019 23 80 2157
25 3649 3002 2774

 

 

94

 

E

+ CBS
+ FT/Al .
+ LB/AIII‘B 2 ;
—C}— MOET/MIPS "
——6— urlFEB .,
wow F820 °

—-— F840 '« ' /

 

 

 

1500

 

S

 

‘é

 

 

 

Genetic progress in milk yields (kg)

 

O
J__

 

t l.. I
0 l 2 3 4 5 6 7 8 9 10111213141516171819 20 2122 23 24 25

Yet

Figure 3-1. Estimated trend of genetic progress of the target cow population under

proposed breeding schemes over 25 years

For F840 and F820 schemes, the semen were imported at year 0 and mated to the

cow population at the same year. The bulls were born 8 year ago and completed ﬁrst crop

progeny test in the US. The progress of genetic merit of females in the population started

at year I and genetic progress of cow population started at year 4, which was 3 years

earlier than that of others proposed breeding.

Comparisons among proposed breeding schemes

The rank on rate of genetic progress after 25 year for the breeding schemes was

STFEB, untFEB, MOET/AI/NS, F820, F 840, untLB/AI/NS, CBS and PT/AI with 2141,

2053,1948,1856, 1771,1667, 1588, and 1339 kg/cow, respectively. All proposed breeding

schemes, except PT/AI scheme, were superior over CBS by 35%, 29%, 23%, 17%, 12%,

and 5%, respectively (Table 3-9). The rank on genetic progress after 25 year for proposed

95

breeding schemes supplemented by importation of 5% foreign pregnant heifers or 5%
foreign yearling heifers remained the same as expected but inﬂuenced on the magnitude
of the genetic progress.

For the proposed breeding schemes with supplement of importation of 5% foreign
pregnant heifers, the genetic progress at year 25 was higher than that of individual
breeding scheme without supplement by from 364 kg (PT/AI) to 287 kg (STFEB). For
the individual breeding schemes with supplement of importation of 5% foreign yearling
heifers, the genetic progress at year 25 was higher than that of individual breeding
scheme without supplement by from 319 kg (PT/AI) to 261 kg (STFEB). The genetic
progress after 25 years for the supplement practice of importation of 5% foreign pregnant
heifers was higher than that for the importation of 5% foreign yearling heifers by from
55kg (PT/AI) and 26 kg (STFEB).

Table 3-10 showed genetic progress after 25 years for individual breeding schemes or
for individual breeding scheme combined with importation of yearling or pregnant
heifers with supplement of F840 to breed 30% or 50%. PT/AI and LB/AI/NS combined
with importation of pregnant heifers with supplement of F820 to breed 50% cows would
result in higher genetic improvement than that of adopting PT/Al or LB/AI/NS without
supplement. But for the individual breeding scheme of MOET/AI/NS, untFEB/AI and
STFEB/AI/NS combined with importation of yearling or pregnant heifers with
supplement of F 820 to breed 30% or 50%, the genetic progress after 25 years would not
result in higher genetic progress than that of adopting MOET/AI/NS, untFEB/AI and

STFEB/AI/NS without supplement.

96

 

 

Table 3-9. The estimated genetic progress of target population at year 25 under different
breeding schemes

 

 

Breeding Genetic mean Difference % of Rank
scheme at year 25 from CBS the CBS
(abbreviation) (kg) (kg)
Control (C) 1588 0 1.00 7
PT/AI 1339 -249 0.84 8
PT/AI +Fng 1703 115 1.07
PT/AI +FyrH 1658 70 1.04
UntLB/AI/NS 1667 79 1.05 5
UntLB/AI/N8+Fng 203 1 443 1 .28
UntLB/AI/NS+FyrH 1989 401 1.25
MOET/ AI/NS 1882 294 1.19 3
MOET/ AI/N8+Fng 2226 638 1.40
MOET/ AI/NS+FyrH 2190 602 1 .3 8
untFEB/AI 2053 465 1.29 2
untFEB/AI+Fng 2348 760 1 .48
untFEB/AI+FyrH 2321 733 1 .46
STFEB/AI 2141 553 1.35 1
STFEB/AI+Fng 2428 840 1 .53
8TFEB/AI+FyrH 2402 814 1 .5 1
F840 1771 183 1.12 5
F840+Fng 2095 507 1.32
F S40+FyrH 2061 473 1 .30
F 820 1856 268 1.17 4
F 820+Fng 2171 583 1.37
F820+FyrH 2139 551 1.35

 

lCBS= Current Breeding Scheme, PT/AI= Progeny Tested bulls for AI,

untLB/AI/NS= Untested Local elite bulls for AI and NS, MOET/A1/NS= Nucleus
population with Multiple Ovulation and Embryo Transfer to produce AI bulls and NS
bulls, untFEB= Untested bulls from imported foreign embryos for AI, STFEB= AI Bulls
from imported embryos tested by paternal half-sibs, F840= The imported semen selected
from the top 40% proven bulls in U.S. FS20= The imported semen selected from the top
20% proven bulls in U.S. +Fng= Importation of foreign pregnant heifers as the
replacement females supplement to A] scheme. +FylH= Importation of foreign yearling

heifers as the replacement females supplement to AI scheme.

MOET/AI/NS was the best strategy of those using local germplasm. LB/AI/NS ranked

the second. At year 25, the estimated genetic progresses of the commercial cow population

for these two schemes were 1882 and 1667 kg respectively and higher 19 % and 5 % than
that of CBS. Table 3-9 showed the genetic change over 25 years in the nuclear population.
MOET/AI/NS scheme had an advantage of an initial genetic lift from the founder stock of
the nucleus herd. In these two schemes, 100 NS bulls were produced by contract mating
and utilized MOET each year and 200 bulls were in natural service to mate 6% of cow
population. The genetic progress at year 25 for the STFEB and FEB strategy were 2246 kg
and 2053 kg respectively, STF EB scheme was better than the untF EB by 9.4%. Yet, there
were more 6000 straws of SB semen imported which were sires of embryos and needed
economic efﬁciency analysis to determined if the genetic difference between STFEB and
FEB lead to greater economic beneﬁts.

Table 3-10. Genetic improvement in milk yield (kg) at year 25 for the alternative
breeding schemes with supplement of F 840 or F820 to breed 30% or 50% cows.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Breeding F 8 F840 F 840 F820 F820

scheme 0% 30% 50% 30% 50%

(abbreviation)

CBS 1588

PT/AI 1338 1484 1573 1515 1622
PT/AI +Fng 1702 1835 1916 1862 1960
PT/AI +FyrH 1658 1794 1877 1822 1922

UntLB/AI/NS 1667 1743 1788 1777 1841
UntLB/AI/N8+Fng 2030 2100 2140 2131 2189
UntLB/AI/NS+FyrH 1989 2060 2102 2092 2151

MOET/ AI/NS 1948 1928 1915 1956 1960
MOET/ AI/NS+Fng 2286 2268 2256 2292 2296
MOET/ AI/NS+FyrH 2251 2233 2220 2258 2262

UntFEB/AI 2053 1968 1912 1994 1954
untFEB/AI+Fng 2348 2272 2221 2295 2259
untFEB/AI+FyrH 2321 2243 2191 2266 2230

STFEB/AI 2141 2030 1956 2055 1998
STFEB/AI+Fng 2427 2328 2261 2350 2299
STFEB/AI+FyrH 2402 2300 2231 2323 2270

 

98

 

Importation of the different genetic quality semen of foreign bulls used in Target from
the top 40% to the top 20% does not had very large difference on the genetic progress at
year 25 (1771kg vs 1856 kg). This was because the differences in the selection response
among the bull percentile groups were not large (Table 3-11).

PT/AI breeding scheme was the only one that the genetic progress at year 25 was
lower that that of CBS. In PT/AI scheme, sampling young bulls were assumed to mate to
80 % of cow population from year 3 to year 6. Newborn females would beneﬁt from
genetic contribution from these sampling young bulls since year 4. Progeny tested AI
bulls were proven at year 6. The offspring of these progeny tested proven bulls would be
born at year 7 and would contribute genetic progress to the cow population at year 10.
Due to long generation interval and lower selection intensity (20%) for the SC, the
genetic progress of cow population was very slow and realized much later than that of
other proposed schemes. At year 25 , only 1339kg improvement was obtained and was
lower 16% than that of CBS. Progeny testing and artiﬁcial insemination has continued to
be the backbone of dairy cattle improvement and currently operate in most developed
countries though it was a time consmning and expensive operation task. Modern PT/AI
schemes lead to substantially higher estimates of annual genetic change than previously
hypothesized and should reﬂect selection intensity increases as well as shortened
generation intervals. Yet, these parameters which affect genetic progress were difﬁcult to
achieve in practice for target population breeding schemes involving imports.

Most of bulls produced from alternative breeding schemes were untested. Superior
young animals needed to be identiﬁed. Parent Average (PA) has been utilized for this

purpose for many years. Though parent average was a good predictor of the bull's

99

eventual proof, some young sires would come through with proofs below their parent
average, while others would come through with higher proofs. The net result was that the
average proofs for a group of bulls was usually quite close to the overall average of the
bulls' parents. Overall, little difference existed between the genetic merit of young sires
and the genetic merit of the Active Al bulls that were available at the same time the
progeny-tested bulls were sampled (Powell and Norman, 1989). The weakness of the PA
as a predictor was the accuracy of predicted transmitting abilities (PTA) for bull-dams.
Though the implementation of the animal model greatly increased the accuracy of cow
PTA, there was still a tendency of overestimating PTA for bull-dams (Ferris et a1. 1991,
Graham et a], 1991, Mac et a], 1991, Uimari et a]. 1992). If bull dams could be reliably
identiﬁed, the use of untested young bulls could be more favorable than progeny testing
(Smith and Burnside, 1990). Thompson and Freeman (1972) found that the use of
untested sons of progeny bulls gave rapid genetic gain in small population.

Hunt et a]. (1974) compared estimates of genetic gain from progeny testing programs
with estimates when the same population was bred only to young bulls except for
contract mating. All young bulls gave similar rates of genetic progress to progeny testing
schemes for the smallest cow population. The all young bulls gave comparable results to
progeny testing when a low percent of the population was recorded and a low percent of
recorded cows were bred to young test bulls. The numbers of young bulls required to
achieve the rates of genetic gain shown were high. Target dairy population was limited
size with 100,000 cows. The untested bulls produced from breeding schemes would use
as a group with conﬁdence for regular service to breed cows, and if bull-dams could be

reliably identiﬁed, the optimum breeding scheme still could be identiﬁed with untested

100

young bulls.

F840 and F 820 could make higher and faster genetic gains because they were from
dairy populations with a genetic mean higher than that of Target. These two schemes
started to contribute genetic progress at year 4 while other schemes started at year 7. At
year 7, these two schemes had contributed genetic progress by 173 kg and 209 kg
respectively. It showed that adopting semen imports strategy in the earlier period and
later switching to a local selection program should contribute more genetic progress.
Semen imports strategy in the earlier period upgraded the domestic stock and then further
improved the new domestic stock through local selection programs. Moreover, imports
would increase genetic variation in the population, which would ensure the success of a

well-run progeny testing scheme.

Table 3-11. The genetic change in each selection path for breeding schemes

 

 

 

Breeding Genetic change
scheme
(abbreviation) SB DB SC DC
Control (C)- untAI (20%) 482 459 232 -
F850(60%) - - 208 -
PT/AI 600 569 538 -
UntLB/AI/NS 600 569 378 -
MOET/ AI/NS - - -
Nucleus population 600 275 484 253
Target population - - 378 -
untFEB/AI - - 214 -
STFEB/AI - - 219 -
F 840 - - 245 -
F820 - - 356 -

 

Selection of parents gives rise to a temporary disequilibrium with less genetic
variation for selection among offspring (Buhner, 1971). For this study we assumed that

these two populations had been under continuous selection for milk yield and had

101

 

 

 

reached selection equilibrium. Therefore the genetic parameters used already partly
include the reduction due to the Bulmer effect. If the assumption does not hold the
Bulmer effect, it would reduce the genetic mean in practice and the population mean
estimated. In progeny testing scheme, the Bulmer effect would reduce response by 21%
and genetic variance by 27% (Meyer and Smith, 1990). The reduction in genetic
variance due to selection was likely to be less in MOET schemes because of less accurate
and less intense selection than progeny testing. In Jeon et a]. (1990), the estimated
reduction in genetic variance in MOET schemes was 17%. However, these values were
dependent on the parameters used. Keeping the population open to introduce germplasm
from the exporting country would result in higher genetic means and reduce the effects of
inbreeding (Hodges, 1991). When calculating the genetic merit mean of new born
animals, we didn't consider the Bulmer effect.
The effect of genotype by environment interaction

The effect of genotype by environment interaction on the genetic progress at year 25
when the initial difference between the U8 and Target population was 603kg was showed
in Table 3-12. Because the initial genetic mean difference between two populations was a
constant and selection response were low for the breeding schemes in this study, the
ranking of schemes did not change when the r8 increases from .5 to .9. But it inﬂuenced
the magnitude of the genetic level of bulls at birth and the genetic progress for a breeding
scheme.

By importing foreign germplasm form North America to a subtropical area, genotype
by environment interaction had been shown not to be important amongst temperate

countries (Petersen, 1975; Garabano et al., 1989; Sieber and Powell, 1989). Yet, the

102

 

results reported for GXE interactions between temperate and tropical countries were
inconsistent and inconclusive possibly due to limited sets of data available. In McDowell
eta]. (1976) report, the genetic correlation estimate for milk production between Mexico
and the USA was high (.86) whereas in Abubakar et a]. (1987) report, it was moderate
(.51). The lowest genetic correlation in performance found in the literature was .08,
between Sri Lanka and Denmark (Buvanendran and Petersen, 1980). In Chauhan
(1983), the genetic correlation between Denmark and India was .11 while that between
USA and India was .7. These ﬁgures showed serious GXE interaction. INTERBULL
estimated of genetic correlations among countries. Estimates would normally range from
.86 to .89 between two North Hemisphere countries and from .75 to .78 between a North
and a South Hemisphere country. The interaction between the genetic correlation of
countries and the initial genetic difference between the importing and the exporting
counties also inﬂuenced ranking of the schemes. Mpofu (1993) reported that when the
initial genetic difference was low and the r8 between the two countries was high (1.0),
genetic progress was faster for all schemes both based on imports and on local selection.
But when the initial genetic difference was high with a low genetic correlation, it meant
that the genetic correlation between countries should be given more weight than the

initial genetic difference between countries when choosing the import countries.

103

 

 

Table 3-12. The effect of different genotype by environment interaction ( rg ) on genetic
improvement of milk yield for the proposed breeding schemes at year 25

 

 

 

Breeding Genotype by environment interaction
scheme
(abbreviation) 1'8 = -5 1'8 =-7 r8 =-9
Control (C) 1539 1588 1638
PT/AI 1260 1339 1416
PT/AI +Fng 1626 1 702 1719
PT/AI +FyrH 1581 1658 1 7 34
UntLB/AI/NS 1577 1667 1757
UntLB/AI/N8+Fng 1943 2030 21 18
UntLB/AI/N8+FyrH 1901 1989 2077
MOET/ AI/NS 1773 1881 1990
MOET/ AI/NS+Fng 2121 2225 2330
MOET/ AI/NS+FyrH 2084 2] 89 2295
untFEB/AI 2006 2053 2099
untFEB/AI+Fng 2301 243 8 2395
untFEB/AI+FyrH 2274 2321 2368
STFEB/AI 2067 2141 2215
STFEB/AI+Fng 2355 2427 2500
STFEB/AI+FyrH 2329 2402 2475
F840 1717 1771 1825
FS40+Fng 2040 2095 2149
FS40+FyrH 2007 2061 21 15
F820 1778 1856 1934
F820+Fng 2095 2171 2247
F820+FyrH 2062 2139 2215

 

CONCLUSIONS AND IMPLICATIONS
There are several breeding schemes used in different dairy populations in the world.
This study studied 7 alternative breeding strategies designed for practical use in the target
population. Because all breeding schemes involved importation of foreign germplasm,
inbreeding depression and genetic drift, which would have otherwise inﬂuenced genetic

changes over generations, were ignored.

104

Among the designed breeding schemes, STF EB scheme with supplement of
importing 5% pregnant heifers had the highest cumulative genetic gain in milk yield at
year 25. In this scheme, imported elite embryos produced bulls as 8C and imported 6000
straws of semen from sires of embryos, which was F810, were also as SC to mate cows
in target to produce paternal half-sibs to test the bulls produced from embryos. Both bulls
from embryos and bulls of sires of embryos contributed genetic progress to the
population. Because frozen embryos to produced AI bulls and semen of SB were
imported from a foreign source, large amount of importation costs imposed on this
scheme. Economic efﬁciency needed to be evaluated to determine if it was an optimum-
breeding scheme for target dairy population. The second optimum scheme was untFEB.
This scheme only imported enough embryos to produce required AI bulls.

Amongst the schemes utilizing local resources, MOET/AI/NS was better than CBS,
PT/Al and LB/AI/NS respectively. Poor reproductive efﬁciency of dairy cows was an
obstacle to genetic improvement because of long generation interval. MOET/AI/NS and
LB/AI/NS produced AI and NS bulls. These two schemes were special tailor-made
breeding schemes for target population situation. NS bulls produced from MOET or
contract matings with known genetic merit and would contribute genetic progress to the
population and improve reproductive efﬁciency. These N8 bulls, however, would only
stay as NS bulls for two years in herd, each herd only with one NS bull. Each bull would
only mate 30 cows a year in a herd. Genetic contribution from them to the population was
only 6%. To produce more NS bulls with known genetic merit and AI bulls with high
genetic merit, MOET and [VP P (In Vitro Fertilization Production) could make it and

would be very important technologies applied to breeding scheme in Target in future

105

study.

Total reliance on foreign semen with the genetic progress at 25 years were higher
than that resulted from CBS. F840 and F820 could make higher and faster genetic gains
because they were from dairy populations with a genetic mean higher than that of target
population. The genetic progress adopting F820 or F 840 schemes showed that the genes
ﬂow from the foreign population to the importing population was faster in SC path than
in SB path.

Scheme PT/AI has been standard in most developed countries. However, in this
study, PT/AI was the only scheme, of which genetic gain was lower than that of CBS.
Lower selection intensity and longer generation interval were the primary reasons.

Ranking of schemes was not affected by the inclusion of a supplement of importation
of either 5% pregnant heifers or 5% yearling heifers, or a supplement of F 820 or F840 to
breed 30% or 50% cows. However, the supplements did affect the magnitude of all
genetic gains after 25 years.

In this study, PTA M was the sole selection criterion for all breeding schemes. Other
traits such as stress resistant traits in fertility, calving ease, mastitis resistance, and other
health traits perhaps need to be considered in breeding goals as well. However, the
development of a total merit index and the establishment of an overall breeding goal are

beyond the scope of this study.

106

 

CHAPTER 4

ECONOMIC EFFICIENCY OF ALTERNATIVE BREEDING SCHEMES
FOR A DAIRY CATTLE POPULATION OF LIMITED SIZE

ABSTRACT

The geographically isolated dairy cattle population of 100,000 in Taiwan population
was the target population for genetic improvement. Eight practical alternative breeding
schemes and schemes with supplement of additional imported germplasm were designed.
In the previous Chapter, these schemes were ranked with respect to their rates of genetic
progress. This study attempted to rank the same schemes for their economic efficiency.
For each of the schemes, appropriate governmental agencies were expected to bear all
costs for the establishment and maintenance of its infrastructure and operation. Farmers
would be expected to pay the costs for semen, AI, imported heifers, as well as increased
costs of feeds, health and fertility problems due to increased milk production.

The economic efﬁciency for a breeding scheme was calculated as the Net Present
Value of its accumulated beneﬁt over a 25 years time horizon. The ranking of the
designed breeding schemes based on their economic efﬁciency was different from that
based on their rates of genetic progress. All schemes involved the importation of pregnant
or yearling heifers resulted in relatively high rates of genetic improvement, but were not
efﬁcient economically. The economic efﬁciency ranking, in descending order, were
STFEB, FEB, MOET/AI/NS, LB/AI/NS, FS40, PT/AI and, F820. All of these alternative
breeding schemes with the exception of the last two, surpassed the current breeding

scheme (CBS) in economic efﬁciency by 7%, 6%, 5%, 3%, and 1%, respectively.

107

 

 

Amongst the schemes utilizing local resources, MOET/AI/NS was better than CB8,
PT/Al and LB/AI/NS by 5%, 6% and 2%, respectively. Amongst all the breeding
schemes designed, the use of imported embryos to produce AI bulls tested by half—sib
performance (STFEB) was the best in terms of economic efﬁciency, as well as, as shown
earlier in the previous Chapter, in terms of the rate of genetic progress. Different levels of
genotype by environment interaction affected the magnitude of estimated economic

beneﬁts but did not affect the ranking of these schemes.

INTRODUCTTON

Milk production is the major source of income for a typical dairy cattle production
system accounting for 80 to 90% of the gross income. Because milk yield is reasonably
heritable, it should be therefore the primary selection goal for all dairymen. The rate of
genetic progress achieved dependents upon the sources of sires and dams and the
breeding scheme used. Early investigations on identifying an optimum-breeding was
giving maximum genetic progress. (Rendel and Robertson,1950). Lindhe (1968) and
Oltenacu & Young (1974) showed that it was not sufﬁcient to characterize a breeding
scheme merely in terms of genetic gains. Miller (1977) reviewed economic studies of
selection programs for artiﬁcial insemination and concluded that there was a diminishing
increase in rates of genetic gain with increasing costs of a breeding program. More
recently, the trend had been to follow up genetic evaluations with economic evaluations
(Ruane and 8mizh, 1989; Dekkers and Shock, 1990; Mpofu etal., 1993). Dekkers and
shook (1990) presented genetic and economic consequences of hybrid nucleus breeding

schemes for AI ﬁrm operating under circumstances in the US Holstein population. Open

108

 

adult schemes increased present value of returns from 20 years of selection by 10% to
20%. Relative gains in returns to progeny-tested bulls increased from 8 to 16% for the
ﬁrst 4 years of selection to 46 to 63% for the ﬁrst 20 yr. They also concluded that large
adult MOET scheme might not be proﬁtable to run.

Costs and returns for a breeding scheme accumulated over time and were realized
over different periods. They must be expressed in comparable time units. The usual
method of comparison was through discounting which future costs and beneﬁts were
transformed to current value. The economic efﬁciency of a breeding scheme was best
measured over a long but ﬁnite period. Gains obtained early were of greater monetary
value than those obtained later in the scheme (Smith, 1981).

An optimum breeding scheme would need to be the one that gives high rate of
genetic progress, but at reasonable costs and maximum beneﬁts to the dairy industry.
Therefore, the speciﬁc objectives of this study were to use Net Present Value (NPV) to
determine the economic beneﬁts and costs of each of the proposed breeding schemes in

Chapter 3, and to re-rank the same schemes based on their economic efﬁciency.

METERALS AND METHODS
Current breeding scheme and proposed alternative breeding schemes
Current Breeding Scheme summarized the real current breeding practices in target
population. It includes 20% of cows mated to local selected but untested bulls by A]
service, 60% of cows mated to imported semen by AI, and 20% of cows mated to
unknown quality bulls by natural service. The genetic change calculated for CBS was

used as the base for comparisons with other alternative breeding schemes.

109

 

Seven proposed alternative breeding schemes and different preportion (30% and
50%) of imported foreign semen combined with alternative schemes and 5% of import
heifers supplemented to AI were studied.

(1) Ofﬁcial progeny tested bulls for AI (PT/AI)

10 progeny tested bulls were from contract mating that the imported semen of F85
mated to top 5 % PTA M cows. Bulls were proven by their 100 daughters' performance.
Semen from these bulls would be collected and stored after sampling. They would mate
80% of target cow population aﬁer proven.

(2) Untested Local elite bulls for AI and NS (LB/AI/NS).

20 AI bulls and 200 N8 bulls were produced from the same contract mating as
PT/AI. 20 AI bulls were selected by the rank of estimated Parent Average (PA) for PTA
M to sire of 80% cows by A1. 200 NS bulls were in 200 herds. Each bull mated to 30
cows in the herd by NS and would contribute to 6% of target population genetic progress.
(3) Nucleus population with Multiple Ovulation and Embryo Transfer to produce AI bulls
and NS bulls (MOET/AI/NS)

20 AI bulls and 200 NS bulls were produced ﬁ'om the controlled Nucleus population
bred by Multiple Ovulation and Embryo Transfer (MOET) techniques. AI bulls were
selected by the estimated Parent Average (PA) for PTA M to sire of 80% cows by AI.
200 NS bulls from MOET were in 200 herds. Each bull mate to 30 cows in the herd by
N8 and would contribute to 6% of target population genetic progress.

(4) Untested bulls from imported foreign embryos for AI (untFEB)
Ten AI untested bulls were produced from imported frozen embryos but transferred

in local recipients. These embryos were from planned mating of elite cows and bulls in

110

 

the exporting country and selected on the estimated Parent Average (PA) for PTA M.
They would mate 80 % of target cow population.
(5) AI Bulls from imported embryos tested by patema] half-sib (STFEB)

Ten AI bulls were produced from imported frozen embryos but transferred in local
recipients. These embryos were from planned mating of elite cows and bulls in the
exporting country. 600 doses semen of the sires of embryos would also imported at the
same time so that there would be 100 patemal half-sib to the bulls. Bulls were selected by
the half-sister performance in the local population.

(6) Total reliance on foreign semen (F 840 and F820)

A total of 227,000 doses frozen semen were imported each year to mate 80% target
cow population. The imported semen were selected from the top 40% and 20 % proven
bulls list on PTA M with 70% REL from the exporting countries to sire target cow
population respectively.

(7) Different proportion (30% and 50%) of imported semen with different genetic quality
(F840 and F820) combined with alternative scheme

For example, /AI+FS40-30% is above alternative breeding schemes with supplement
of F 840 breed 30% of the target cow population. Four combinations were study. They
were /AI+FS40-30%, /AI+FS40-50%, /AI+F820-30%, /AI+FS20-50%.

(8) Individual breeding schemes supplemented by additional imports

Two kinds of additional imports supplemented to individual breeding schemes. One
is semen importation from t0p 20% or 40% to breed 30% or 50% of the target cow
population. The second is 5% heifers imported supplement to individuals. AI/+Fng

stands for importation of foreign pregnant heifers as 5% of the replacement females

1]]

 

supplement AI schemes. AI/+FylH stands for importation of foreign yearling heifers as
5% of the replacement females supplement AI schemes.
The model of evaluation- NPV

The economic beneﬁts of the genetic improvement from a breeding scheme were
determined by (1) The initial cost of germplasm, (2) Price of milk produced, (3) Cost of
milk production and extra costs of feeds and health and fertility problems due to
increased milk production, (4) Discount rates, and (5) Time horizon.

Many authors had used discounted cash ﬂow method to assess the value of livestock
improvement schemes, and to compare alternative breeding schemes
(Hill.1971,Hink.1971, Petersen.1974, Cunningham and Ryan.1975, Everett.1975). In
Brascamp (1973) and F ewson (1989), beneﬁt was discounted at a higher rate, 10% vs
8%. The justiﬁcation for using higher discount rate was that it allowed for risk. Mpofu
et a] (1993) reported using a high discount rate (10%) to account for the risk of failure of
an investment program because beneﬁt would be recouped far beyond the 25-year
horizon. However, high rates (8 to 15%) used in discounted cash ﬂow analysis of genetic
improvement schemes had tended to underestimate the value of the beneﬁt and to favor
breeding schemes with short-term beneﬁt (Bird and Mitchell.1980, Smith.1978). In this
study, investments and the infrastructure of breeding schemes did not be concerned and
truncating beneﬁt at the end of the evaluation period rather than counting them in
perpetuity, the risk could be ignored and the discount rates could be lower. 80 the
discount rate used in this study was 5% for the economic efﬁciency evaluation. This rate

was recommended for most animal breeding programs (Dekkers and Shook.1990).

112

 

 

 

Therefore, the expected beneﬁt from a breeding scheme would be calculated by
discounted Net Present Value (NPV) of accumulated beneﬁts from the achievable genetic
progress in milk at year 25. Beneﬁts ﬁom the achievable genetic progress in milk were
net return of additional milk due to genetic improvement over feeds and health cost
subtracted the breeding cost.

The Net Present Value (NPV) of the accumulated beneﬁts for a breeding scheme was
calculated as

T
NPV = ZKRBSU) " CBS(r))/(1+ ‘0']

t=l

T
= z[(BBS(r))/(l +61)?

Where I: the evaluation period (year 0 to year 25), which was 26 years,

d: the discount rate 5%,

Rm): the total yearly net return of milk yield from a breeding
scheme at year t,

CBS”): the total costs paid by farmers for semen from bulls
produced from the proposed breeding scheme at year t, and

Easy»: the value of the beneﬁt for farmers from a breeding scheme
at year t

The economic efﬁciency of a proposed breeding scheme was determined by the
difference of the discounted NPV of accumulated relative beneﬁts at year 25 between an

alternative breeding scheme (8433a,) and the current breeding scheme.

NPV - of - relative - beneﬁt - to - CBS

1‘
= Z [(BABSU) — 303.9(1)) /(1 + (1),]
(=1

113

 

 

If NPV difference of accumulated beneﬁt at year 25 between an altemative-breeding
scheme (BABsu) ) and current breeding scheme (BcBS(t)) was negative, it meant the
altemative-breeding scheme was not better than the current breeding scheme. Positive
NPV difference meant alternative breeding scheme was economically viable. The NPV
was also used to rank the proposed breeding schemes.

Calculating NPV of beneﬁt

Table 4-1 showed the average total costs of milk production and beneﬁt per hundred
kg for Target population dairy farmers in 1997. These data were obtained from the ﬁle
of Council of Agriculture. The costs were representative of typical commercial dairy
farms in Target population. Milk production costs were the costs incurred by overheads
which include real estates, buildings, fencing, electricity, and dairy equipment, tractor,
feeds, veterinary services, vet supplies, shipping, selling charges, insurance, basic
overhead expenses (e.g. living expenses of farmers) and miscellaneous costs. Fer the
production cost of extra milk yield due to genetic improvement, only the feed costs and
health-fertility costs were considered. In Table 4-1, the feed costs would be 50% of milk
price per kilogram. Health-fertility costs would be 4% of milk price per kilogram. Here,
the production cost for extra milk from genetic improvement would be taken as 40 % of
milk price per kilogram.

A summary of values was given in Table 4-2 for calculating beneﬁt from the
proposed breeding scheme. They were based on the market price in 1997 in target
population. In this study, milk price was ﬁxed at 1997 market price which was $.61 per
kg. When the supply of milk increases due to genetic improvement, the market price of

milk might decrease because of an increase in the consumer surplus. The price of milk

114

 

 

was determined by the demand for dairy products. The demand was related to consumer
growth. In target population, the consumer population is still low and is growing
especially in young people. The growth in supply and in demand would not lead the price
milk to drop.

Table 4-1. The total costs of milk production and beneﬁt per hundred kg for target
population dairy farmers in 1997

 

 

 

 

 

Items 8 %

£9.81
Feeds 31.38 50.20
Vets and Medicines 2.76 4.42
Labor 5.33 8.53
Insurance .64 1.02
Electricity .64 1 .03
Depreciation of cow 3.88 6.20
Building maintenance .71 1.18
Tool and machine depreciation .70 1.12
Machine maintenance .38 .60
Miscellaneous .09 .15
Artiﬁcial Insemination plus semen price 1.13 1.80

Total cost of 100 kg of milk yield 47.44

Income
Market milk price in Target population In .61

1997

Total Income of 100 kg of milk yield 61

Beneﬁt of 100kg milk yield 13.56 22.23

 

Source: Cost-beneﬁt analysis of milk production for dairy farmers in 1997

Only beneﬁts from milk yield were considered. With the subsiding by government,
farmers would derive full beneﬁt of genetic improvement. Cost for a breeding scheme to
the farmers were the price of semen per straw domestic or imported plus the AI fee.
Assume the price of semen produced from the breeding scheme established by
government would be 8 7 per straw. Table 4-2 showed the costs of milk production and

beneﬁt per hundred-kg milk for target population dairy farmers.

115

Table 4-2. Annual breeding costs to farmers for the proposed breeding schemes

 

 

 

Item $/unit Units Total
($X103)
--_--.C9.$t .......................................................................................................
CBS
F 810 30 142 4
F 850 12 170,100 2,041
Semen from local untested bulls 6 56,700 340
.......... Breeding.f9.9.f9.r.19.c_hnisiarr---__-----------_-----_------_19------2-2.1.9999------.2..1.6.9.----
PT/AI
F85 50 226 11
Semen of bulls produced from scheme 6 227,000 1,362
......... 35991118.£9.9f91399111119129-------------------------------.19-------“2.1.93999------.2.r1.6.9.----
LB/AI/NS
F85 50 850 42
Semen of bulls produced from scheme 6 227,000 1,362
......... Freedmafeafar.Refresher--.-------_--_--_-----------.1.Q------2-2.1.6999---_-_.2_.!.6.9.----
MOET/AI/NS
F85 50 850 42
Semen of bulls produced from scheme 6 227,000 1,362
......... Breeding.terrier.399111119129-----------------_-------_-----1.9---------.2.1.§,QQQ------.2.,.1_6.9.----
untF EB
Semen of bulls produced from scheme 6 227,000 1,362
......... Breeding.teeter.399111319189-------------------------_----19---------.2.1.§,QQQ---2.2.499"---
STFEB
F810 30 6,000 180
Semen of bulls produced from scheme 6 221,000 1,326
......... Breeding.£99-f9_r.399b.r119.ian-__--------_----------------_--1.9-----_--.22.15999"-_----.2.1.9.----
F 840
F840 15 227,000 3,405
......... Breeding£6.9f9.r..t991.1.rris=_iarl--------------------_-----_----19---------.2.1._6.,QQQ------.2..1.6.9.----
F820
F820 20 227,000 4,540
......... Breeding.£ee-f9.r..teehlris=.iarl-.-----------------------------.1.Q-----.---2.1..6.,.QQQ----.-.2..1.6.9.----
/AI+Fng
......... ImpartesllzregaaatheifersEmmi-.3.----------------3.,.3.QQ------------.5.,QQQ-----1§.§.QQ.----
/AI+FylH
Imported yearling heifers from US 2,800 5,000 16,500
Return
Return over additional milk due to genetic .36 Genetic -
improvement per kg change
Imported heifer salvage when sold as beef 800 5,000 4,000

 

116

 

The net return of milk per kg for the extra milk yield due to genetic improvement
would be 60 % of milk price ($.36/kg). So the total yearly net return (RABsm )from milk
yield in the cow population at year t for a breeding scheme would be

RABS(t) = $US .36/kg * genetic progress at year t * 65000 milking cows.
The total yearly costs (CABsm) for a breeding scheme were total paid by farmers for
the price of semen plus the AI fee, which would be
Chasm = (Price of semen + AI fee) * 80,000 * 2.7 * 1.05
where price of semen: the market price of semen of bulls produced from
breeding or cost of imported foreign semen.
80,000: the number of cows bred by A1.
2.7: the AI services per conception.
1.05: for contingency cost allowing for AI technology
failure.
For importation of foreign pregnant or yearling heifers, farmers would pay costs of

imported heifers. It was a kind of investment and had salvage values when the heifers

were sold as beef. These salvage values would be included in the economic efﬁciency as

beneﬁt.

RESULTS AND DISCUSSIONS
Table 4-3 gave the genetic progress in milk yield, discounted cost, income, beneﬁt,
and NPV (Net Present Value) of beneﬁt for each breeding scheme supplemented by
importing foreign pregnant heifers or yearling heifers,. As shown was NPV of relative
beneﬁt from breeding scheme compared with that of CBS at year 25 at the discount rate
5%. The genetic progress were estimated when the initial genetic difference and genetic
correlation between the US and Target population dairy population were 623kg and .7,

respectively. Schemes with positive NPV of relative beneﬁt were economically viable. It

117

 

 

meant the proposed scheme could be used to replace CBS. The ranking of the proposed
breeding schemes for NPV of relative beneﬁt differed greatly from the ranking for
genetic progress in milk yield because of large differences in costs among the schemes,
as detailed in Appendix B Table 1 to 6.

Among these seven proposed breeding schemes, only four of the proposed breeding
schemes which were STFEB, untF EB, MOET/AI/NS, LB/AI/NS had large positive NPV
of relative beneﬁt over 25 years and were economically viable with realistic genetic
progress. F 840 would be economically viable at year 25 but NPV of relative beneﬁt was
negative before year 20.

PT/AI and F820 were negative NPV of relative beneﬁt. PT/AI was negative NPV of
relative beneﬁt at year 25 but was positive NPV before year 20. With discounting, more
beneﬁt realized in earlier years for the PT/AI. NPV of relative beneﬁt for PT/AT was
positive in earlier years but decreasing gradually over years and was eventually $US -
1.06X106 at year 25 because of lower genetic progress than that of CBS. For the
importation of foreign semen schemes (F820 and F840), F840 would be economically
viable at year 25 but NPV of relative beneﬁt was negative before year 20. F 820 was
negative NPV of relative beneﬁt over all 25 years but was gradually close to zero with
increasing over time. There were cumulative high costs involved in running F820 and
F840 schemes at current commercial cost. The F 820 and F840 schemes ranked the fourth
and the ﬁﬁh using genetic progress as a criterion. Yet these two schemes ranking were
the last second and last third. The result was in agreement with study on economic
beneﬁts on farm level for semen importation for Colombia, Mexico and Venezuela by

Holmann et a]. (1990). Average economic beneﬁts were negative for the three countries.

118

 

 

Table 4-3. Economic evaluation for individual scheme at year 25.

 

l

 

 

 

 

 

 

 

Scheme Genetic Cost Return Beneﬁt Accumulated NPV at year 25
progress in Deviation
milk yield Beneﬁt from % of CBS
CBS
kg---- (8x106) ------ % ------
CBS
- ___________ 1.5-83- 1 32 - - 26.0-1- - _ 2.6.6.3- - _ 9.49.7.2 ...... 0 ______ 1 .09- _-
PT/AI 1338 1 07 24.66 23 59 939 67 -l 06 99
+Fng 1703 5 88 28.10 22 22 815.29 -125 43 87
-tEy-IH _______ 16.513 - § .12 - - -21-3.6- - - 22.7.1- - - £17.12. - :23.§1- .. - - _ - .99 - - -
LB/AI/NS 1667 1 08 26.43 25 35 967 50 26 77 103
+Fng 2030 5 89 29.86 23.97 844 30 -96 43 90
-tF .115 ....... 1.9.82- § .21. - - -29-6.4. - - .26-4.2- - - -37-3.6.3- - 35.7.1.9. _ - - - - .93 - .. -
MOET/AI/NS 1948 1 07 27.94 26 87 986 57 45 85 105
+Fng 2286 5 88 31.23 25.35 862 10 -78 63 92
e381 _______ 2 .231. _ 3.21 - _ -31-0.5. _ _ .26-8.6- _ - 39.2.0.3- - refs-6.9. - - - - - .96 - _ -
untFEB 2053 1.07 28.50 27.43 1001.03 60.31 106
+Fng 2348. 5.88 31.57 25.69 871.48 -69.24 93
.rrxlli--_-_-------_.2.3.2.1_--.5..2.1 ....... 3. 1.4.2 ....... 2.6-2.1 ......... 9 93.5.7 ..... -. 3.7-1.6-.------.--.96. .....
STFEB 2246 1.06 29.54 28.48 1007.24 66.52 107
+Fng 2523 5.87 32.50 26.64 877.76 -62.97 93
ﬁfxlﬂ----__----_-_.2..5.QQ--_5..2.9 ....... 3 2.3.3 ....... 2.7-1.9. ........ 9 9.3.4.2 ..... -. 3.2.3.1 ............. 9 .7. .....
F840 1771 1 67 26.99 25 31 941 40 0 68 101
+Fng 2095 6 44 30.20 23 76 815 93 -124 80 87
.213-111.-------_-----.2.Q6..1.--.5_§.1 ....... 3 9.9.2 ....... 2.6-2.1. ........ 6.4.36.6 ..... -. 9.7-2.6. ............ 9 9. .....
F820 1856 2.01 27.44 25.44 935.80 -4.93 99
+Fng 2171 6.76 30.61 23.86 810.37 -130.35 86
+FylH 2139 6.14 30.44 24.30 837.12 -103.61 89

 

ICBS= Current Breeding Scheme, PT/AI= Progeny Tested bulls for AI, untLB/AI/NS=
Untested Local elite bulls for Al and NS, MOET/AI/NS= Nucleus population with
multiple Ovulation and Embryo Transfer to produce AI bulls and NS bulls, UntFEB=
Untested bulls from imported foreign embryos for AI, STFEB= AI Bulls from imported
embryos tested by paternal half-sibs , FS40= The imported semen selected from the top
40% proven bulls in U.S., FS20= The imported semen selected from the top 20% proven
bulls in U.S., +Fng= Importation of foreign pregnant heifers as the replacement females
supplement to A] scheme. +FylH= Importation of foreign yearling heifers as the
replacement females supplement to AI scheme.

For LB/AI/NS, NPV of relative beneﬁt was with low increasing rate over years and

was positive at year 25 .

119

MOET/AI/NS was better than CBS, PT/Al and LB/AI/NS by 5%, 6% and 2%,
respectively. In Dekkers and Shock (1990a) work, the juvenile MOET schemes were
better than progeny testing. F ewson (1989) found that proﬁts were almost the same for
progeny testing and a MOET scheme, those of the progeny testing scheme being 3%
higher. In this study, the nucleus population of MOET/AI/NS scheme was an open one.
Bull Sires were selected from foreign. The nucleus population was expected to be a lot
better genetically than the target since it were better kept open to migration from foreign
populations.

The scheme ranking ﬁrst through third were STFEB, FEB and MOET/AI/NS and
were the same as that of genetic progress. In this study, the genetic advantage of STFEB
and FEB over MOET/AI/NS was the use of foreign bull dams. The genetic means at year
25 for STFEB was higher than that for MOET/AI/NS by a factor of 15% while the
relative NPV for STFEB was higher than that of MOET/AI/NS by a factor of 2%. The
genetic means at year 25 was higher for STFEB than for FEB by a factor of 4% while the
relative NPV for STFEB was higher than that of FEB by a factor of 1%. The genetic
means at year 25 was higher for FEB than for MOET/AI/NS by a factor of 5.3% while
the relative NPV for FEB was higher than that of MOET/AI/NS by a factor of 1%. Use
of imported embryos to produce AI bulls was ranked as the best (STFEB) and the second
best (FEB) of all of the proposed breeding schemes, surpassing CBS, respectively, by 7%
and 6%.

Total reliance on foreign semen (F840 and F820) to mate target population was not
economic viable scheme. But different proportion (30% and 50%) of imported semen

with different genetic quality (F840 and F820) combined with alternative scheme in

120

Table 4-4 were Showed economic viable. Yet, the difference of NPV of beneﬁt from that
of CBS was less than that of individual breeding schemes except PT/AI. PT/AI combined
with importing foreign semen at different proportion become economic viable with the
most when combined with F S40-30%.

None of those schemes supplemented by importing foreign pregnant or yearling
heifers was economically viable alternatives. They were not viable even though each
heifers imported had a salvage value of $800 in addition to their contribution to genetic
progress to the population.

Genetic correlation between two populations did not affect signiﬁcantly the ranking
of schemes based on economic efﬁciency evaluation in Table 4-5 but had an effect on the
absolute values of the NPV of relative beneﬁt because rg affected the magnitude of
genetic progress. PT/AI scheme became economic viable when rg was 0.9. This result
was agreed to that of Mpofu et a1 (1993). They reported when no interaction in
performance occurs between countries, the importation strategies have higher NPV, but
the ranking of proposed breeding schemes was little changed. When the interaction
increased (rg became less), the local testing strategies improved relatively in their NPV.

Economic efﬁciency analysis help to identify economical viable schemes and the
optimum breeding scheme based on whether it would contribute signiﬁcantly to the
development of the economy or beneﬁt to farmers. The study showed that STFEB was
an optimum-breeding scheme in terms of genetic improvement and economic efficiency
for Target population dairy cattle population. The cost of STFEB was list in Table 4-6 for
bulls selection program. If this scheme was adopted by the government agencies, a bull

station needed to be established and experienced technicians for frozen embryos transfer

121

 

 

Table 4-4. Economic evaluation for individual schemes with supplement of F840 or
F820 to breed 30% or 50% cow population at year 25

 

 

 

 

 

 

 

 

 

 

Breeding Genetic Cost Income Beneﬁt NPV deviation
Scheme 1 progress in from CBS
milk yield
---- kg ---- ($X106)

CBS 1588.68 1.37 26.01 24.63 0
PT/AI 1338.78 1.07 24.66 23.59 -1.06
+FS40-30% 1521.38 1.21 25.64 24.42 4.06
+FS40-50% 1580.76 1.31 25.96 24.64 3.61
+F820-30% 1546.10 1.29 25.77 24.47 2.87
+F820-50% 1620.51 1.45 26.17 24.72 1.57
LB/AI/NS 1667.27 1.08 26.43 25.35 26.77
+FS40-30% 1729.12 1.22 26.76 25.53 23.02
+FS40-50% 1766.33 1.32 26.96 25.64 20.32
+F820-30% 1757.12 1.30 26.91 25.60 22.16
+F820-50% 1810.63 1.45 27.20 25.74 18.78
MOET/AI/NS 1948.97 1.07 27.94 26.87 45.85
+FS40-30% 1932.94 1.21 27.85 26.63 37.01
+FS40-50% 1922.20 1.31 27.79 26.48 31.13
+F820-30% 1954.65 1.29 27.97 26.67 35.84
+F820-50% 1958.45 1.44 27.99 26.54 29.18
untF EB 2053.13 1.07 28.50 27.43 60.31
+F840-30% 1985.32 1.21 28.13 26.82 45.99
+FS40-50% 1940.25 1.31 27.89 26.58 36.45
+FS20—30% 2005.35 1.29 28.24 26.95 44.65
+F820-50% 1974.24 1.44 28.07 26.63 34.21
STFEB 2246.05 1.06 29.54 28.48 66.52
+FS40—30% 2052.32 1.19 28.49 27.30 45.44
+FS40-50% 1993.65 1.28 28.18 26.89 36.61
+F820-30% 2073.43 1.27 28.61 27.34 44.40
+F820-50% 2027.63 1.40 28.36 26.95 34.88

 

lCBS= Current Breeding Scheme, PT/AI= Progeny Tested bulls for Al, untLB/AI/NS=
Untested Local elite bulls for AI and NS, MOET/AI/NS= Nucleus population with
multiple Ovulation and Embryo Transfer to produce AI bulls and NS bulls, UntFEB=
Untested bulls from imported foreign embryos for AI, STFEB= AI Bulls from imported
embryos tested by paternal half-sibs , FS40= The imported semen selected from the top
40% proven bulls in U.S., F820= The imported semen selected from the top 20% proven
bulls in U.S., +Fng= Importation of foreign pregnant heifers as the replacement females
supplement to AI scheme.+FylH= Importation of foreign yearling heifers as the
replacement females supplement to AI scheme.

122

 

 

Table 4-5. The effect of genotype by environment interaction on NPV of relative
beneﬁt to that of CBS from the proposed breeding schemes at year 25 .

 

 

 

 

 

 

 

Breeding NPV of relative beneﬁt to CBS from scheme
schemel rlg =.5 rg =.7 r8 =.9
(8x106)

CBS 0 0 0
PT/AI -2. l 3 -l .06 0.02
UntLB/AI/NS 24.85 26.77 28.72
MOET/ AI/NS 42.01 45.85 49.32
untFEB/AI 59.14 60.31 61.47
STFEB/AI 57.21 58.70 60.02
F840 -0.63 0.68 2.03
F 820 -9.53 -4.93 -0.37

 

lCBS= Current Breeding Scheme, PT/AI= Progeny Tested bulls for AI,
untLB/AI/NS= Untested Local elite bulls for AI and NS, MOET/AI/NS= Nucleus
population with multiple Ovulation and Embryo Transfer to produce AI bulls and NS
bulls, UntFEB= Untested bulls from imported foreign embryos for AI, STFEB= AI Bulls
from imported embryos tested by paternal half-sibs , FS40= The imported semen selected
from the top 40% proven bulls in U.S., F S20= The imported semen selected from the top
20% proven bulls in U.S.,

Table 4-6. Yearly costs for bull station to adopt STFEB scheme

 

 

Items Calculation 8 (X103)
Bull station investment 312,500 3 12.50
Items replacement per year 312,500*0.05 15.62
Importing frozen embryos 1,000*56(8TFEB) 56.00
Hormone for synchronizing recipientsl 40*56embryos*4 8.96
Bull maintenance costs 2,800*10 bulls 28.00
Young bull maintenance costs 2,100* 14 bulls 29.40
Yearling bulls maintenance 1,700*10 bulls 17.00
Semen collection, storage, and delivery 2*227,000 straws 454.00
Total 921.00
Overheads for bull station per year 609.00
Income for semen sale 7*227,000 straws 1,589.00

 

 

1 Four heifers were prepared for one embryo transferring.
were employed by the government. These two factors related to the success of this
scheme. The cost in Table 4-6 did not include the salary of these technicians. The bull

station got income from sale of semen, which were $1,589,000 a year. The amount could

123

cover the overheads of bull station.

Mpofu et a] (1993) proposed breeding strategies for genetic improvement of dairy
cattle in Zimbabwe. The strategy that ranked best economically was the continuous
importation of elite embryos to generate young. untested bulls, which was like as untF EB

in this study.

IMPLICATIONS AND CONCLUSIONS

This study proposed 8 alternative breeding strategies and several combinations of
them, all were practical for application in the target population for genetic improvement
for milk yield. Among these proposed breeding schemes, STFEB was a practical and the
optimum breeding scheme in terms of both the rate of genetic progress and economic
efﬁciency. In this scheme, imported elite embryos would produce bulls as sire of cows
(SC) and imported 6000 straws of semen from Sires of embryos, which was F810, as SC
to mate cows in the target population to produce paternal half-sibs to test the bulls
produced from embryos. Both bulls from embryos and bulls of sires of embryos
contributed genetic progress to the population.

The second optimum scheme was untF EB. This scheme did not import F810 to mate
cows and the number of embryos was less and the cost of this scheme was cheaper than
STFEB. Yet, more genetic gain in milk yield could cover the extra cost and appeared to
be the optimum breeding scheme.

MOET/AI/NS scheme was a local selection scheme and was the third optimum
scheme. Schemes utilizing local resources such as MOET/AI/NS and LB/AI/NS would

also result in higher genetic progress in milk yield at 25 years than that resulted from

124

 

 

CBS and be also economic viable. These two schemes were special tail-made alternative
breeding schemes for target dairy situation. NS bulls produced from MOET in nucleus
population or contract matings with known genetic merit would contribute genetic
progress to the population and improve reproductive efﬁciency too. For the nucleus
population, beneﬁts would also come from sale of these NS bulls that were not included
in the evaluation.

Total reliance on F 820 was not economically viable due to expensive import cost
though it result in higher genetic gain. F820 ranked forth based on genetic improvement,
but ranked the last based on economic efﬁciency. F 840 was economically viable. It
implicated that the imported semen didn't need to be ﬁom the best bulls in the exporting
country. The most important was that the bulls should be of higher genetic merit in
exporting country than top 40% progeny tested proven bulls in the importing country.

PT/AI has been the backbone scheme of dairy cattle improvement in most developed
countries. In this study, PT/Al was the only scheme which genetic gain and NPV of
relative beneﬁt was lower than those of CBS. The rank of PT/AI was the last based on
genetic improvement, but still ranked near the bottom based on economic efﬁciency.

Importation of F 820 or F840 to breed 30% or 50% cows combined with LB/AI/NS
would result in higher genetic improvement and less NPV of beneﬁt. But MOET/AI/NS,
untFEB, and STFEB with supplement of importation of F820 or F840 to breed 30% or
50% cows, the genetic improvement were lower and NPV of beneﬁts were less than
those of individual schemes without supplement.

Individual breeding schemes with supplement of importation of 5% pregnant heifers

or 5% yearling heifers resulted in the magnitude of the genetic improvement higher than

125

 

 

 

those without supplement. But they all were not economic viable.

The results of this study indicated that foreign elite bulls and foreign elite dams of
bulls must be more effectively used because of the great initial genetic difference
between the target and the U.S. dairy cattle populations. Even the genetic by environment
interaction existed, effective use of foreign elite bulls and foreign elite dams of bulls still
could make maximal genetic progress. Some of the other alternatives within each of
schemes and combined with other schemes might make them more efﬁcient. Imports of
high quality foreign semen from elite bulls to mate only higher performance local cows
with some proportion might be economic viable breeding practice. We didn't use the
females from embryos as bull-dams. The genetic merit of these females from embryos
Should be as elite as the bulls from the embryos. They should be used as bull-dams in

PT/AI, LB/AI/NS and even MOET/AI/NS scheme, especially in the earlier years.

126

 

 

CONCLUSIONS

Sale of milk is the major source of income for most dairy producers. Milk yield,
therefore, Should receive heavy emphasis in selection programs. The dairy population
around the world is decreasing in size. This fact means genetic improvement in milk
yield is continuous increasing through breeding and the dairy producers must become as
efﬁcient and as proﬁt-minded as possible.

Milk yield was closely related to economic merit. The genetic response in milk
yield would steadily be increasing keeping on selection with various currently available
selection criteria from USDA and Holstein Association USA. When milk was paid for its
volume with a fat differential or with both fat and protein differential, selection on PTA
M could get largest response in milk yield with small difference on MF$, MFP$, NM$.
These milk yield selection criteria would also increase SCS in milk while much less from
selection on NM$. Linear type traits such as rear leg of Side view, rump angle, fore udder
attachment, udder depth and teat placement would be resulted in undesirable responses.
These undesirable responses probably be the cause the health and fertility stress
problems. If the severity of these undesirable responses increased, the value of increased
milk yields would be offset by the additional cost for these undesirable responses.

When farmers set up breeding goal to breed a more proﬁtable herd without concern
about cow shape, they might use yield or yield merit or NM$ as selection criterion. They
select AI sires with high value of these selection criteria and then scrutinized heavily for
PTA SCS and for linear type traits such as fore udder attachment and udder depth. If

farmers wanted to change cow conformation to meet their personal standards, selection

127

on TPI or additional selection pressure on linear type traits after screening on yield
selection criteria would be a better sire selection practice.

There is a variety of breeding schemes designed for dairy industry in the world.
This study cited 8 alternative breeding strategies and schemes with additional supplement
of imported germplasm which were all practical under target population conditions.
Because all breeding schemes involved importation of foreign germplasm, the factors,
which inﬂuenced genetic gains such as inbreeding depression and genetic drift, are
ignored.

Among these proposed breeding schemes, STFEB scheme could result in highest
genetic gain in milk yield at year 25. In this scheme, imported elite embryos produced
bulls as SC and imported 6000 straws of semen ﬁ'om sires of embryos, which was F810,
were also as SC to mate cows in target to produce paternal half-sibs to test the bulls
produced from embryos. Both bulls from embryos and bulls of Sires of embryos
contributed genetic progress to the population. A large amount of importation costs
imposed on this scheme. Yet, economic efﬁciency evaluation Showed that more genetic
gain in milk yield could cover the extra cost and obtained more beneﬁts at year 25.

STFEB appeared to be the best optimum practical breeding scheme.

The second optimum scheme was untFEB. This scheme did not import F810 to

mate cows and the number of embryos was less and the cost of this scheme was cheaper

than STFEB. The genetic progress at 25 years of total reliance on foreign semen was
higher than that from CBS. F840 and F820 could make higher and faster genetic gains
because they were from dairy populations with a genetic mean higher than that of target
population. The genetic progress adopting F820 or F840 schemes showed that the genes

ﬂow from the foreign population to the importing population was faster in SC path than

128

 

in SB path. Economic efﬁciency evaluation showed that F820 were not economic viable
due to expensive import cost though it result in higher genetic gain. F840 was economic
viable. It implicated that the imported semen didn't need to be from the best bulls in the
exporting country.

Amongst the schemes utilizing local resources, MOET/AI/NS was better than CBS,
PT/Al and LB/AI/NS respectively. Poor reproductive efﬁciency of dairy cows was an
obstacle to genetic improvement because of long generation interval. MOET/AI/NS and
LB/AI/NS produced AI and NS bulls. These two schemes were special tailor-made
alternative breeding schemes for Taiwan dairy situation. NS bulls produced ﬁom MOET
or contract matings with known genetic merit and would contribute genetic progress to
the population and improve reproductive efﬁciency. These NS bulls, however, would
only stay as NS bulls for two years in herd, each herd only with one NS bull. Each bull
would only mate 30 cows a year in a herd. Genetic contribution from them to the
population was only 6%. To produce more NS bulls with known genetic merit and AI
bulls with high genetic merit, MOET and IVFP could make it and would be very

important technologies applied to breeding scheme in Taiwan in future study.

PT IAI was the only scheme which genetic gain and NPV of relative beneﬁt was
lower than those of CBS. Lower selection intensity and long generation interval limited
the genetic gain. It is unlikely that all individual breeding schemes were adopted alone.

PT /AI combined with all other proposed schemes such MOET/AI/NS, untFEB, and
STFEB, and even proportion of F820 or F840 should be resulted in higher genetic

improvement and more economical efﬁciency.

The results of this study implied that target dairy population Should be more

effective use of foreign elite bulls and foreign elite dams of bulls as SB or DB because of

129

the high initial genetic difference between the target and US dairy cattle population. Even
the genetic by environment interaction existed, effective use of foreign elite bulls and
foreign elite dams of bulls still could make maximal genetic progress. Some of other
alternation within each of schemes and combined with other schemes might make them
more efﬁcient. Imports of high quality foreign semen from elite bulls to mate only higher
performance local cows with some proportion might be economic viable breeding
practice. The use of females from embryos as DB in PT/AI, LB/AI/NS and even
MOET/AI/NS scheme especially in the earlier years should be more effective.

When genotype by environment interactions exist, the best genotype in one
environment was not the best in another. Using Taiwan dairy cattle milk yield data ﬁles
to evaluate the genetic correlation between Taiwan and US population was also a
important study in the recent future.

In this study, only milk yield was of interest. Because of speciﬁc resources,
management system and environment in Taiwan, other traits of dairy cows such as type
traits, stress resistant traits, fertility, calving ease, mastitis resistance, and even general
health Should be monitored when improvement focus on milk yield. To deve10p a total
merit index with economic weights is important for the future studies. Future study
should move towards improvement in multi-trait breeding goal but not reduce milk yield

much.

130

 

 

APPENDIX A

131

APPENDIX A

Table 1. Average breeding value by birth year
for milk yield of the U.S. Holstein.

 

 

Year Cows BV Sire BV
57 31273 -5069 -4918
60. 137257 -5001 -4827
65 188330 -4625 -4381
70' 247104 -4101 -3682
75 315750 -3300 -2675
80. 540206 -2286 -1425
85. 615421 -1208 -374
86 638114 -986 -137
87 663625 -767 74
88 679817 -516 339
89 692542 -237 651
90 702666 0 880
91 703594 279 1197
92 680665 571 1515
93 651727 838 1791
94 656233 1100 2061
95 635800 1321 2270

 

AIPL - USDA report (1998,FEB)

132

APPENDIX A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

——x —- PT/Al bulls y
400° 1“ —-—s-— PTIAI Females ;
,- 3500 ~4 —x— PT/Al+Hng females , A. 3494
g 3000 . —<>—— PTIAI+Fy4H females x w, 5:: /" 2472
re A ’ r
“5’ 2500 ”Re/r "‘ 5675‘: 2550
'2 x 1‘ 56:5 1,4 i 2227
2 2m x ’X x "x % ﬁf—l
8 1500 ,. new gig-22*: .r
x =4 ’4‘ 7x / (52:2? " 3
1000 , e652
/ x2995?”
500 __ gogQ '“ i
/ 4)...-
0 «rover-care . . e . . . . r a 4
o 2 4 6 8 10 12 14 16 18 20 22 24
Year
2500 ﬂ --4:l—- PT/Al+Fng
—o— PTIAl+FyIH ,
2000 .51658
17
1500 seer"; 03

W l 1339
500 i
0 " Ff! I r r r r T r r r r r r r r r r r 1

0 2 4 6 81012141618 20 22 24
Years

Genetic progress in milk yield (kg)

 

 

Figure 1. Estimated genetic trend of breeding stock populations and genetic progress of
the target cow population for PT/AI 1 and supplement by F ng2 and FylH3

' PT/AI= progeny tested bulls for AI

2 Fng= scheme with supplement of imported pregnant heifers
3 Fy1H= scheme with supplement of imported pregnant heifers.

133

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

we“ NS bulls _
4000 . —o—- Al bulls ;
3500 + 131811820? a... 3‘82
‘ "0— pg em A :
a ——<>— LB/All +1: IH 3'03
:3: 3000 NS y females /./:.: 32905
t .
g 2500 =
.9
3; 2000
C
3 1500
1000
500
0 1
3000 . - - _ ~—~ ~ --.-_ ~-
3 —<>—— LB/Al/hS .
5 -—o-— LB/Al/NS+F H
g 2500 44 —o—- LB/AI/lB+F)i ;
3 $1989
~5- E1667
lg 1500 . f 3
O. 1000 - l
o p.
.3 .
8 ﬂ :
(3 500 '-
0 «rs-1+: ﬂﬁ/X r f r

Figure 2. Estimated genetic trend of breeding stock populations and genetic progress of
the target cow population for untLB/AI/NS and supplement by Fng2 and F ylH3

' untLB/AI/NS = Untested Local elite bulls forAI and NS

2 Fng = scheme with supplement of imported pregnant heifers
3 FylH = scheme with supplement of imported pregnant heifers.

134

4500 «

 

  
 
 
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 +018 bulls
400 + Al bulls
a 3500 _..... MOET/Al/NS females
3“: 3000 —o——MOET/AIINS++bgH females
E. 2500 —o— MOET/Al/NS+FyI-I females
0
a; 2000
C
8 1500
1000 -
500 .

0
2‘6: --<>-— MOET/N/NS
4; —o— MOETlN/NS+FyIl-l 2226
g 20m 2190
E lﬂ/j/‘Z/ 1882
.s
3’3 1500 ’
e .
3’
o. 1000
.2
E W
8 500 M

O 3r'rVTV—rvmclvr r r r r r ﬁT‘r ﬁr f

 

Years

Figure 3. Estimated genetic trend of breeding stock populations and genetic progress of
the target cow population for MOET/AI/NS and supplement by Fng2 and FylH3

' MOET/AI/N8= Nucleus population with Multiple Ovulation and Embryo Transfer to
produce AI bulls and NS bulls,

2 F ng= scheme with supplement of imported pregnant heifers

3 F ylH= scheme with supplement of imported pregnant heifers.

135

4500- ~ = m-mw ~ ,-
+Al bulls

4000 ‘ —+— untFEB females .
3500 - —o— untFEB+Fng females / 3624

3000 ——o— untFB-kFyll-I females / 3.1”” 32:;
25m // {ﬁf‘ .
2000 / £36241:

1500 r r"

1000 146?:

0 2 4 6 8 10 12 14 16 18 20 22 24

 

 

 

 

 

Genetic merit (kg)

 

 

 

 

 

m"... .. ..,..-..-. .,... .. . ..,.....-...... .. .. .... .. _............-_....-..... .. . ......-.. ..,..-.-.. .. .. .. .. ...... ....~~..._...»_ .. .. .. -.M..............-.a.....a..—.~....... .. .....~—.........

 

3000 -. ---
—<>— untFEB

—~o- untFEB+F H
2500 -~ -—<>— untFEB+Fyl

 

 

 

 

2000 / g 2031
,1667
1500 W/ i

1000

5,, W

 

 

Genetic progress in milk yield (kg)

 

Figure 4. Estimated genetic trend of breeding stock populations and genetic
progress of the target cow population for untFEB/AI ' and supplement by Fng2
and F ylH3

1 untFEB/AI= Untested bulls from imported foreign embryos for AI

2 Fng= scheme with supplement of imported pregnant heifers
3 FylH= scheme with supplement of imported pregnant heifers.

136

 

4500- ._ _ _ . - _ - -

 

 

   

 

 

 

 

 

 

 

 

 

 

 

+" ' “Al bulls“ ;
4000 -~ —-e—STFEB females *3745
3500 .. —D—STFEB+Fng females / M 3698
a -——+- STFEB+Fy4ermales .3 ,a/ 33350
e5 3000 .
s
e 2500
'3 2000
a)
I:
1WD . 3.: . 4,",
500
O‘Mrrrrrrrrrrrrrrrrrtrﬁtﬁ

 

 

 

2500 “0"“ ”=ng
—o— +FylH M 92141
1500

1 000 W
500

 

 

 

Genetic progress in milk yield (kg)

 

Figure 5. Estimated genetic trend of breeding stock populations and genetic progress of
the target cow population for STFEB/AI ' and supplement by Fng2 and FylH3

l STFEB/AI= AI Bulls from imported embryos tested by paternal half-sibs

2 Fng= scheme with supplement of imported pregnant heifers
3 FylH= scheme with supplement of imported pregnant heifers.

137

_e- - _._ _ -. .. ..-__ - -..__ ........7

4000

     
 

+ F840 Al bulls

 
 
    

 

 

—o— F840+Fng females 3174
a 3000 —<>— FS40+FyH females
é ; 2874
.95 2500
‘e’
2 2000
ES
5 1500
1000
500
0
0 2 4 6 8 1O 12 14 16 18 20 22 24
Year
3000 . .
—+— F840 Al bulls
—-o— FS40+Fng
2500 .
+ FS4°+WH 2095

 

 

 

2000 / 2061
1771

1500 -/

1000 W

,,, ﬂ

0 2 4 6 8 10 12 14 16 18 20 22 24
Year

Genetic progress in milk yield (kg)

 

 

Figure 6. Estimated genetic trend of breeding stock populations and genetic progress of
the target cow population for F840 1 and supplement by Fng2 and FylH3

I FS40= The imported semen selected from the top 40% proven bulls in U.S.

2 Fng= scheme with supplement of imported pregnant heifers
3 FylH= scheme with supplement of imported pregnant heifers.

138

4000

 

3500 <
3000 i

 

——0— F320 N bulls
* F820 females

 

 

—0— FSZO+Fng females

 
 

 

 

—<>— FS20+FylH females

 

 

 

 

 

Genetic merit (kg)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

g 3000 —o—FSZO
g 2500 g —o— F820+Fng 1
; ~a—- FSZO+FylH 2171
= 22150
5 2000 M ?
.s :1856
2 1500 u , f
,3 1000
5 //
C
8 500 M
0 :. .: ,,,,, , , , , , , , T f f T

Figure 7. Estimated genetic trend of breeding stock populations and genetic progress of

Year

14

20 22 24

the target cow population for F 820 l and supplement by Fng2 and FylH3

' F820= The imported semen selected from the top 40% proven bulls in U.S.
2 F ng= scheme with supplement of imported pregnant heifers
3 FylH= scheme with supplement of imported pregnant heifers.

139

APPENDIX B

140

APPENDIX B

Table 1. Economic efﬁciency evaluation for individual schemes using local bull-dams
over 25 years

 

 

 

 

 

 

 

Year Genetic Cost Income Beneﬁt NPV of NPV
progress in (2) (2) (I) (2) beneﬁt difference
milk (3) of beneﬁt
from CBS
(3)
---kg--- $x106
CBS
0 0.00 3.74 59.15 55.41 55.41 0
5 37.33 3.64 46.88 43.23 293.74 0
10 319.00 2.85 39.88 37.02 490.07 0
15 701.76 2.24 34.60 32.36 660.90 0
20 1131.97 1.75 30.06 28.31 810.35 0
25 1588.68 1.37 26.01 24.63 940.72 0
PT/AI
0 0.00 0.01 59.15 59.14 59.14 3.74
5 0.00 2.85 46.35 43.50 306.24 12.50
10 213.99 2.23 38.70 36.47 499.95 9.88
15 594.49 1.75 33.66 31.91 668.41 7.52
20 945.10 1.37 28.78 27.41 814.29 3.94
25 1338.78 1.07 24.66 23.59 939.67 -1.06
LB/AI/N S
O 0.00 0.04 59.15 59.11 59.11 3.70
5 0.00 2.87 46.35 43.48 306.08 12.34
10 343.37 2.25 40.15 37.90 503.14 13.07
15 751.88 1.76 35.03 33.28 679.24 18.34
20 1207.15 1.38 30.57 29.19 833.24 22.89
25 1667.27 1.08 26.43 25.35 967.50 26.77
MOET/AI/N S
0 0.00 0.00 59.15 59.15 59.15 3.74
5 0.00 2.84 46.35 43.50 306.28 12.53
10 350.28 2.23 40.23 38.00 503.25 13.17
15 893.76 1.75 36.28 34.53 683.38 22.49
20 1424.68 1.37 32.07 30.70 844.62 34.27
25 1948.97 1.07 27.94 26.87 986.57 45.85

 

(1) Income — Cost

(2) Discounted value
(3) Discounted accumulated value

141

Table 2. Economic efﬁciency evaluation for individual schemes using imported

germplasms over 25 years

 

 

 

 

 

 

 

Year Genetic Cost Income Beneﬁt NPV of NPV
progress in (2) (2) (I) (2) beneﬁt difference
milk (3) of beneﬁt
from CBS
(3)
---kg--- $x10‘5
untFEB
O 0.00 0.00 59.15 59.15 59.15 3.74
5 0.00 2.84 46.35 43.50 306.28 12.53
10 429.56 2.23 41.11 38.88 505.56 15.48
15 1001.69 1.75 37.22 35.48 690.45 29.55
20 1535.44 1.37 32.83 31.46 855.89 45.53
25 2053.13 1.07 28.50 27.43 1001.03 60.31
STFEB
0 0.00 3.60 59.15 55.55 55.55 0.15
5 2.99 2.82 46.39 43.57 296.13 2.38
10 476.71 2.21 41.64 39.43 496.83 6.76
15 1112.45 1.73 38.19 36.46 686.02 25.13
20 1691.37 1.36 33.89 32.54 856.76 46.41
25 2246.05 1.06 29.54 28.48 1007.24 66.52
F S40
0 0.00 5.67 59.15 53.48 53.48 -1.93
5 54.55 4.44 47.12 42.68 286.03 -7.71
10 391.89 3.48 40.69 37.21 482.32 -7.75
15 816.88 2.73 35.60 32.88 655.10 -5.79
20 1284.24 2.14 31.10 28.97 807.63 -2.72
25 1771.38 1.67 26.99 25.31 941.40 0.68
F320
0 0.00 6.80 59.15 52.35 52.35 -3.06
5 68.69 5.33 47.33 41.99 280.25 -13.49
10 447.06 4.18 41.31 37.13 475.25 -14.82
15 890.05 3.27 36.24 32.97 648.25 -12.64
19 1268.84 2.69 32.55 29.85 772.30 -9.75
20 1365.58 2.56 31.66 29.10 801.39 -8.96
25 1856.50 2.01 27.44 25.44 935.80 -4.93

 

“’ Income — Cost

(2) Discounted value

(3) Discounted accumulated value

142

 

 

Table 3. Economic efﬁciency evaluation for individual schemes using local bull-dams
supplemented by importing pregnant heifers over 25 years

 

 

 

 

 

 

 

Year Genetic Cost Income Beneﬁt NPV oi NPV
progress (2) (2) (”(2) beneﬁt difference
in milk (3) of beneﬁt
from CBS
(3)
---kg--- 39x106
PTAI+Fng
O 20.21 16.51 64.52 48.01 48.01 -7.40
5 98.32 15.60 51.67 36.07 250.24 -43.50
10 407.98 12.22 43.94 31.72 415.37 -74.71
15 848.85 9.58 38.29 28.71 564.97 -95.92
20 1254.10 7.50 32.78 25.28 698.17 -112.18
25 1703.00 5.88 28.10 22.22 815.29 -125.43
LBAI/AI/NS+ F ng
0 20.21 16.54 64.52 47.98 47.98 -7.43
5 104.46 15.62 51.75 36.14 250.35 -43.39
10 547.32 12.24 45.50 33.26 419.44 -70.64
15 1008.03 9.59 39.68 30.09 576.80 -84.09
20 1518.35 7.51 34.59 27.08 718.28 -92.07
25 2030.98 5.89 29.86 23.97 844.30 -96.43
MOET/AI/NS+ Fng
0 20.21 16.50 64.52 48.02 48.02 -7.39
5 104.46 15.59 51.75 36.16 250.55 -43.20
10 555.67 12.22 45.59 33.37 419.57 -70.51
15 1143.48 9.57 40.87 31.29 580.91 -79.99
20 1717.63 7.50 35.96 28.46 729.08 -81.27
25 2286.12 5.88 31.23 25.35 862.10 -78.63

 

(I) Income — Cost

(2) Discounted value

(3) Discounted accumulated value

143

 

Table 4. Economic efﬁciency evaluation for individual schemes using imported

germplasms supplemented by importing pregnant heifers over 25 years

 

 

 

 

 

 

 

Year Genetic Cost Income Beneﬁt NPV of NPV
progress in (2) (2) (I) (2) beneﬁt difference
milk (3) of beneﬁt
from CBS
(3)
---kg--- 9006
untFEB+ Fng
0 20.21 16.50 64.15 47.65 47.65 -7.76
5 98.32 15.59 51 .67 36.07 249.65 -44.09
10 620.34 12.22 46.31 34.10 420.37 -69.70
15 1223.56 9.57 41.57 32.00 585.32 -75.57
20 1792.43 7.50 36.47 28.97 736.44 -73.91
25 2348.47 5.88 31.57 25.69 871 .48 -69.24
STFEB+ Fng
0 20.21 19.87 64.52 44.65 44.65 -10.76
5 101.18 15.57 51.71 36.14 240.77 -52.97
10 666.51 12.20 46.83 34.63 412.89 -77.19
15 1326.21 9.56 42.47 32.91 581.90 -78.99
20 1934.42 7.49 37.45 29.96 737.88 -72.48
25 2523.03 5.87 32.50 26.64 877.76 -62.97
F S40+ F ng
0 20.21 21.82 64.52 42.70 42.70 -12.71
5 150.60 17.09 52.41 35.32 231.33 -62.42
10 567.97 13.39 45.73 32.34 398.69 -91.38
15 1049.97 10.49 40.05 29.56 552.06 -108.84
20 1564.47 8.22 34.91 26.69 691.27 -119.09
25 2095.30 6.44 30.20 23.76 815.93 -124.80
FSZO+ Fng
O 20.21 22.88 64.52 41.64 ‘ 41.64 -13.77
5 164.03 17.93 52.60 34.68 225.92 ~67.83
10 618.92 14.05 46.30 32.25 392.05 -98.02
15 1116.26 11.01 40.63 29.62 545.54 -115.35
20 1637.53 8.62 35.41 26.79 685.21 -125.14
25 2171.48 6.76 30.61 23.86 810.37 -130.35

 

(1) Income — Cost

(2) Discounted value

(3) Discounted accumulated value

144

 

 

Table 5. Economic efﬁciency evaluation for schemes using local bull-dams
supplemented by importing yearling heifers over 25 years

 

 

 

 

 

 

Year Genetic Cost Income Beneﬁt NPV of NPV
progress (2) (2) (I) (2) beneﬁt difference
in milk (3) of beneﬁt
from CBS
(3)
---kg--- $X106
PT/AI+ FylH
0 0.00 14.01 64.15 50.14 50.14 -5.26
5 85.65 13.64 51.48 37.85 262.43 -31.31
10 386.41 10.69 43.70 33.01 434.84 -55.23
15 822.82 8.37 38.06 29.69 589.96 -70.93
20 1217.49 6.56 32.53 25.97 727.14 -83.21
25 1658.11 5.14 27.86 22.71 847.12 -93.61
LB/AI/NS+ FylH
O 0.00 14.04 64.15 50.11 50.11 -5.30
5 91.00 13.84 51.56 37.73 261.93 -31.81
10 528.57 10.84 45.29 34.45 437.56 -52.51
15 982.83 8.49 39.46 30.97 599.96 -60.94
20 1485.28 6.66 34.37 27.71 745.04 -65.31
25 1989.33 5.21 29.64 24.42 873.63 -67.10
MOET/AI/NS+ FylH
O 0.00 14.00 64.15 50.15 50.15 -5.26
5 91.00 13.81 51.56 37.75 262.13 -31.62
10 537.49 10.82 45.39 34.57 437.69 -52.38
15 1123.86 8.48 40.70 32.22 604.24 -56.66
20 1690.62 6.64 35.77 29.13 756.23 -54.12
25 2251.62 5.21 31.05 25.84 892.03 -48.69

 

(1) Income — Cost

(2) Discounted value

(3) Discounted accumulated value

145

Table 6. Economic efﬁciency evaluation for schemes using imported

germplasms supplemented by importing yearling heifers over 25 years

 

 

 

 

 

 

 

 

 

Year Genetic Cost Income Beneﬁt NPV of NPV
progress (2) (2) (I) (2) beneﬁt difference
in milk (3) of beneﬁt
from CBS
(3)
---kg--- $x106
untFEB+FylH
0 0.00 14.00 64.52 50.52 50.52 -4.89
5 85.65 13.81 51.48 37.67 262.53 -31.22
10 606.67 10.82 46.16 35.34 439.95 -50.12
15 1209.65 8.48 41.45 32.97 610.37 -50.53
20 1771.79 6.64 36.33 29.69 765.54 -44.81
25 2321.09 5.21 31.42 26.21 903.57 -37.16
STFEB+FylH
0 0.00 17.60 64.15 46.55 46.55 -8.85
5 88.59 13.79 51.53 37.74 251.75 41.99
10 654.56 10.80 46.70 35.89 430.62 -59.46
15 1315.82 8.46 42.38 33.91 605.24 -55.66
20 1917.93 6.63 37.33 30.70 765.42 -44.93
25 2500.46 5.20 32.38 27.19 908.42 -32.31
FS40+FylH
0 0.00 19.67 64.15 44.48 44.48 -10.93
5 139.37 15.41 52.25 36.84 241.65 -52.09
10 550.19 12.08 45.53 33.45 415.29 -74.79
15 1029.52 9.46 39.87 30.41 573.47 -87.42
20 1537.29 7.41 34.72 27.31 716.22 -94.13
25 2061.39 5.81 30.02 24.21 843.46 -97.26
FSZO+FylH
0 0.00 20.80 64.15 43.35 43.35 -12.06
5 153.20 16.30 52.45 36.15 235.88 -57.87
10 602.54 12.77 46.11 33.34 408.09 -81.98
15 1097.69 10.01 40.47 30.46 566.30 -94.59
20 1612.23 7.84 35.24 27.40 709.44 -100.91
25 2139.45 6.14 30.44 24.30 837.12 -103.61

 

(1) Income - Cost
(2) Discounted value
(3) Discounted accumulated value

146

 

BIBLIOGRAPHY

 

147

BIBLIOGRAPHY

Allaire, RR, and JP. Gibson. 1992. Genetic value of herd life adjusted for milk production.
J. Dairy Sci.75zl349.

Banos,G., and GE. Shook. 1990. Genotype by environment interaction and genetic
correlations among parities for somatic cell count and milk yield. J. Dairy Sci. 73 :2563

Bichard, M., A.H.R. Pease, P.H. Swales, and K.Ozkutuk. 1973. Selection in a population
with overlapping generations. Anim. Proc.17:215.

B1ake,R.W., F.J.Holmann, J.Gutierrez, and GF. Cevallos. 1988. Comparative
proﬁtability of United States artiﬁcial insemination sire in Mexico. J. Dairy Sci.
71:1378.

Boettcher, P.J. L.K. Jairath, K.R Koots and J .C.M. Dekkers. 1997. Effects of interactions
between type and milk production on survival traits of Canadian Holsteins. J. Dairy
Sci. 80:2984-2995.

Boettcher, P.J., L.B. Hansen, H. Chester-Jones, and CW. Young. 1993. Responses of
yield and conformation to selection for milk in a designed experiment with a control
population. J. Dairy Sci. 76:267-273.

Boettcher, P.Reneau, J.,L.B.Hanse, P.M.VanRaden, and C.A.Ernst. 1992. Genetic
evaluation of Holstein bulls for somatic cells in milk of daughter. J. Dairy Sci. 75:1127

Boldman, K. G., and T. R. Famula. 1985. Association of sire dystocia transmitting ability
with progeny linear type traits in Holsteins. J. Dairy Sci. 68:2052.

Boldman, K.G., A. B. Freeman, B. L. Harris, and A. L. Kuck. 1992. Prediction of sire
transmitting abilities for herd-life from transmitting abilities for linear type traits. J.
Dairy Sci. 75: 552-563.

Boldman, KG., A.E. Freeman, B.L.Harris, and AL. Kuck. 1992. Prediction of sire
transmitting ability for herd life from transmitting ability for linear type traits. J. Dairy
Sci. 75:552

148

Brascamp, B.W., J .A.M. van Arendonk, and AF.Groen. 1993. Economic appraisal of the
utilization of genetic markers in dairy cattle breeding. J. Dairy Sci. 76:120~1214.

Burnside, E.B., G.B.Tanson, G.Civati, and E.Dadati. 1992. Observed and theoretical
genetic trends in a large dairy population under intensive selection. J. Dairy Sci.
75:2242.

Burke,BP., and D.AFunk. 1993. Relationship of linear type traits and herd life under
different management system. J. Dairy Sci. 7622773.

Cassell,B.G. 1986. Maxbullzcomputerized sire selection for the herd. NCDHIP
HANDBOOK.

Cassell,B.G., RE. Pearson, J.Stoel, and S.Hiemstra. 1990. Relationships between sire
evaluations for linear type traits and lifetime relative income ﬁ'om grade and registered
daughters. J. Dairy Sci.73: 198.

Chang,C.C., C.Y. Tseng, Z.Y.Chen, and MC. Chen. 1997. The Holstein Diary Herd
Improvement program in Taiwan. Taiwan Livestock Res.Vol.30 (1). 55-65.

Coffey,E.M., W.E.Vinson, and R.E.Pearson. 1986. Potential of somatic cell
Concentration in milk as a sire selection Criterion to reduce mastitis in dairy cattle. J.
Dairy Sci. 69:2163.

Colleau J.J. 1991. Using embryo sexing within closed mixed multiple ovulation and
embryo transfer schemes for selection on dairy cattle. J .Dairy Sci. 74: 3973-3984.

Colleau,J.J. and Le Ban Duval, E. A simulation study of selection methods to improve
mastitis resistance of dairy cows. J. Dairy Sci. 78 : 659-671.

Cue, R.J., and J. F. Hayes. 1985. Correlations of various direct and maternal effects for
calving ease. J. Dairy Sci. 68:374.

Dadati,F.B., B.W. Kennedy, and EB. Bumude.1985. Relationships between
conformation and reproduction in Holstein cows: type and calving performance. J.
Dairy Sci. 68:2639.

149

De Haan, M.H.A., B.G. Cassell, R.E. Pearson , and BB. Smith. 1992. Relationship
between net income, days of productive life, productive life, production, and linear
type traits in grade and registered Holstein. J. Dairy Sci.75z3553.

Dekkers, J.C.M. and Shook, GE. 1990. Genetic and economic evaluation of nucleus
breeding schemes for commercial artiﬁcial insemination ﬁrms. J .Dairy Sci. 73:1920-
1937.

Dekkers,J.C.M. 1993. Theoretical basis of genetic parameters of herd life and eﬂ'ect on
response to selection. J. Dairy Sci.76:1433.

Dekkers,J.C.M., L.K. Jairath, and EH. Lawrence. 1994. Relationship between sire genetic
evaluations for conformation and functional herd life of daughters. J. Dairy Sci.77:844.

Dekkers, C. M. 1994. Optimal breeding strategies for calving case. J. Dairy Sci. 77: 3441
—3453

Dekkers.J.C.M.l992. Structure of breeding programs to capitalize on reproductive
technology for genetic improvement. J. Dairy Sci. 75:2880-2891.

Dekkers, J.C.M. and GE. Shook.1990. Economic evaluation of alternative breeding
programs for commercial artiﬁcial insemination ﬁrms. J. Dairy Sci. 73:1902-1919.

Dekkers, J.C.M. and GE. Shook.1990. Genetic and economic evaluation of nucleus

breeding schemes for commercial artiﬁcial insemination ﬁrms. J. Dairy Sci. 73:1920-
1937.

Dentine,M.R., B.T. McDaniel, and H.D.Norman. 1987. Evaluation of sires for traits
associated with herdlife of grade and registered Holstein cattle. J. Dairy Sci. 70:2623.

Ducroco,V. and Quaas,R.L. 1988. Prediction of genetic response to truncation selection
across generations. J .Dairy Sci. 71 :2543-2553.

Ducrocq, V.,R.L. Quaas, E.J. Pollak, and G. Casella. 1988. Length of productive life of
dairy cows. 2. Variance component estimation and sire evaluation. J.Dairy Sci. 71 :3071.

150

Emanuelson,U., B. Danell, and J. Philipsson.1988. Genetic parameters for clinical mastitis,

Somatic cell counts ,and milk production estimated by multiple-trait restricted maximum
likelihood. J. Dairy Sci.7lz467.

Everett, R.W. 1975. Income over investment in semen. J. Dairy Sci. 58:1717.

Everett, R. W., J. F. Keown, and E. E. Clapp. 1976. Relationships among type,
production, and stayability in Holsteins. J. Dairy Sci. 59:1505.

Everett, R.W. 1984. Impact of genetic manipulation. J. Dairy Sci. 67:2812-2818.

Falconer, D. S. 1989. Introduction to quantitative genetics. 3rd ed. Academic Press,
Longman, Hong Kong.

Foster, W. W., A. B. Freeman, and P. J. Berger. 1988. Linear type trait analysis with
genetic parameter estimation. J. Dairy Sci. 7t:223,

Henderson, C. R. 1976. A simple method for computing the inverse of a numerator
relationship matrix used in prediction of breeding values. Biometrics 32: 69.

Heuven,H.C.M., H. Bovenhuis, and RD. Politiek. 1988. Inheritance of somatic cell count
and its genetic relationship with milk and composition in Holstein. J. Dairy Sci. 65:843.

Hill, W.G. 1971. Investment appraisal for national breeding programs. Anim. Prod.
13:37.

Hill, W.G. 1974. Prediction and evaluation of response to selection with overlapping
generations . Anim. Prod. 1 8: 1 17.

Holmannf., R.W. Blake, R.A. Milligan, R.Barker,P. A. Oltenacu, and M. V. Hahn. 1990.
Economic returns from United States artiﬁcial insemination sires in Holstein herds in
Colombia, Mexico, and Venezuela. J. Dairy Sci. 73:2179.

Hopkins, I.R., and J.W. James. 1977. Some optimum selection strategies and age
structures with overlapping generations. Anim. Prod. 25:111.

151

Hopkins, LR. and J .W. James. 1979. Genetic responses in the early years of selection
programmes using genetic differences between generations. Anim. Prod. 28:65.

Jairath, L. ,J.C.M. Dekkers, L.R. Schaeffer, Z. Liu, E.B. Burnside and B. Kolstad 1998.
Genetic evaluation for herd life in Canada. J. Dairy Sci. 81 :550-562.

Juga, J. and Maki-Tanila, A. 1987. Genetic change in a nucleus breeding dairy herd using
embryo transfer. Acta, 37:511- 519.

Keller, D. S. and Teepker. 1990. Effect of variability in response to superovulation on
donor cow selection differentials in nucleus breeding schemes. J. Dairy Sci. 73:549

Kennedy,B.W., M.S. Sethar, J.E. Moxley, and BK Downey. 1982. Heritability of somatic
cell count and its relationship with milk yield a composition in Holsteins. J.Dairy Sci.
65:843.

Koops, W. J., M. Grossman, and J.H.G. den Daas. 1995. A model for reproductive
efﬁciency of dairy bulls. J. Dairy Sci. 78:921.

Kuhn, M.T. ,P.J. Boettcher and A.E. Freeman.l994. Potential biases in predicted
transmitting abilities of females from preferential treatment. J. Dairy Sci. 77:2428-
2437.

Lee, K. L., A. B. Freeman and L. P. Johnson.1985. Estimation of genetic trend in the
registered Holstein cattle population. J. Dairy Sci. 68:2629.

Lindhe, B. 1968. Model simulation of AI breeding within a dual purpose breed of cattle.
Acta Agric. Scand. 18:33.

Jeon, G.J., I.V. Mao, j. Jensen, and T. A. Ferris. Stochastic modeling of multiple
ovulation and embryos transfer breeding schemes in small closed dairy cattle
population. J. Dairy Sci. 73:193 8-1944.

Louis, M.M., C. Smith and J.C.M. Dekkers. 1993. MOET results from a dispersed
hybrid nucleus programme in dairy cattle. Anim. Prod. 57:369-378.

152

Mao. I.L., M.C. Dong and CE. Meadows. Selection of bulls for progeny testing using
pedigree indices and characteristics of potential bull-dam’ erds. 1991. J. Dairy Sci. 74:
2747.

Meuwissen, T.H.E. 1991. Expectation and variance of genetic gain in open and closed
nucleus and progeny testing schemes. Anim. Prod. 53:133.

Meuwissen,T.H.E., and J .A.M. Van Arendonk. 1992. Potential improvements in rate of

genetic gain from marker- assisted selection in dairy cattle breeding schemes. J. Dairy
Sci. 75: 1651- 1659.

Meyer, K., and C. Smith. 1990. Comparison of theoretical and simulated equilibrium
genetic response rates with progeny testing in dairy cattle. Anim. Prod. 50: 207.

Misztal, I., T.J. Lawlor, T. H. Short, and P. M. VanRaden. 1992. Multiple-trait estimation
of variance components of yield and type traits using an animal model. J. Dairy Sci.
75:544.

Mocquot J.C., 1988. Sire-Son and darn paths: research geneticist's view. J. Dairy Sci.
71:1972 -1981.

Monardes, H. G., R. I. Cue, and J. F. Hayes. 1990. Correlations between udder
conformation traits and somatic cell Count in Canadian Holstein cows. J. Dairy Sci.
73:1337.

Mopofu, N., Smith, C. and Burnside. E. B. 1993. Breeding Strategies for genetic
improvement of dairy cattle in Zimbabwe. l.genetic evaluation J .Dairy Sci. 76:1163-
1172.

Mopofu, N., Smith, C., Van Vuuren, W. and Burnside, E. B. 1993b. Breeding strategies
for genetic improvement of dairy cattle in Zimbabwe. 2. Economic evaluation. J.
Dairy Sci. 76:1173-1181.

Nicholas, F .W., and C. Smith. 1983. Increased rates of genetic change in dairy cattle by
embryo transfer and splitting. Anim. Prod. 36:341.

153

Norman, H. D., and L. D. Van Vleck. 1972. Relationship of ﬁrst lactation production and
type traits with lifetime performance. J. Dairy Sci. 55: 1726.

Rendel J. M.and Robertson . 1950. Estimation of genetic change in milk yield by
selection in a close herd of dairy cattle. J. Genetic 50: 1-8.

Rogers. G.W., and B.T. McDaniel. 1989. The usefulness of selection for yield and ﬁmctional
type traits. J. Dairy Sci. 72:187.

Rogers,G.W., B.T.McDaneil, M.R.Dentine, and D.A.Funk. 1989.Genetic correlation
between survival and Linear type traits measured in ﬁrst lactation .j Dairy Sci. 72:523.

Rogers,G.W., G.L. Hargrove, T.J. Lawlor,Jr., and J.L. Ebersole.l991. Correlations among
linear type traits and somatic cell count counts. J. Dairy Sci.74:1087.

Rogers. G.M.,B. T. McDaniel, M. R Denteine, and LP. Johnson . 1988. Relationship
among survival rates, predicted differences for yield, and linear type traits. J. Dairy Sci.
71:214.

Rogers, G.W. 1993. Index selection using milk yield, somatic cell score, udder depth, teat
placement, and foot angle. J.Dairy Sci. 76:664

Ruane, J ., and J. J. Colleau. 1996. Marker-assisted selection for a sex-limited character
in a nucleus breeding population. J. Dairy Sci. 79:1666-1678.

SAS® User's Guide: Statistics, Version 6.03 Edition. 1997. SAS Inst, Inc., Cary, NC.

Schaeffer,L.R., and J .Jamrozik. l 996. Multiple-trait prediction of lactation yields for dairy
cows. J. Dairy Sci. 79:2044-2055.

Schaeffer,L.R., R.Reents and J.Jamrozik. 1996. Factors inﬂuencing international
comparisons of dairy sires. J. Dairy Sci. 79:1108-1116.

Schutz, M.M., L.B. Hansen, G.R. Steuemagel, and J.K. Reneau. 1990. Genetic parameters
for somatic cells, protein and fat in milk of Holsteins. J. Dairy Sci. 73 :494

154

Schutz, M.M., L.B. Hansen, G, R. Steuemagel, and A.L.Kuck. 1990. Variation of milk, fat,
and somatic cells for dairy cattle. J Dairy Sci. 73:484.

Schutz, M.M.1994. Genetic evaluation of somatic cell score for United State dairy cattle. J.
Dairy SCi. 77:2113.

Schutz, M.M., P.M. VanRaden, P.J. Boettcher, and L. B. Hansen. 1993. Relationship of
somatic cell score and linear type trait evaluation of Holstein sires. J. Dairy Sci. 76:284.

Schutz,M.M., and P.M.VanRaden, and G.R.Wiggans..l994. Genetic variation in lactation
means of somatic cell scores for six breeds of dairy cattle. J.Dairy Sci.77z648.

Schutz,M.M., P.M.Vanraden, G.RWiggans, and H.D.Norman. 1995. Standardization of
lactation means of somatic cell scores for calculation of genetic evaluations. J. Dairy
Sci.78: 1 843.

Seykora, A.J., and BI. McDaniel. 1986. Genetics statistics and relationships of teat and
udder traits, somatic cell counts, and milk production. J. Dairy Sci. 69:2395

Shook, G. E. 1989. Selection for disease resistance. J. Dairy Sci. 72:1349.

Shook,G.E., and M.M.Schutz.1994. Selection on Somatic cell score to improve resistance to
mastitis in Unite States. J. Dairy Sci. 77:648

Short, T.H., and T.J. Lawlor. 1992. Genetic parameters of conformation traits, milk yield,
and herd life in Holsteins. J. Dairy Sci.75:]987.

Sieber,M., A.F.Freeman, and RN. Hinz. 1987. Factor analysis for evaluating
relationships between ﬁrst lactation type scores and production data of Holstein dairy
cows. J. Dairy Sci. 70:1018.

Sivanadian,B., M.M. Lohuis and J .C.M. Dekkers. 1998. Expected genetic responses from
selection indexes for Canadian dairy cattle under present and future milk pricing
systems. Can. J. Anim. Sci. 78:157-165.

155

Smith, C. 1978. The effect of inﬂation and form of investment on the estimated value of
genetic improvement in farm livestock. Anim. Prod. 26:101.

Smith,C., and J. Ruane. 1987. Use of sib testing as a supplement to progeny testing to
improve the genetic merit of commercial semen in dairy cattle. Can. J. Anim. Sci.
67:985.

Smith C., and M Quinton.1993. The effect of selection in sublines and crossing on
genetic response and inbreeding. J .Anim.Sci. 71: 2631-8

Spelman,R.J., and D.J.Gairick. 1997. Utilization of marker assisted selection in a
commercial dairy cow population. Livest. Prod. Sci. 47:139-147.

Spelman,H.J., and DJ. Garrick. 1998. Genetic and economic responses for with-in family
Marker-Assisted Selection in dairy cattle breeding schemes. J. Dairy Sci. 81:2942-
2950.

Standberg,E., and GE. Shock. 1989. Genetic and economic responses to breeding program
that consider mastitis. J .Dairy Sci. 72.2136.

Stanton,T.L., R.W. Blake, and RL. Quaas.l991. Response to selection of United States
Holstein Sire in Latin America. J.Dairy Sci.74z651.

Strandberg E., E Shock. 1989. Genetic and economic responses to breeding programs
that consider mastitis. J .Dairy-Sci. 72: 2136-42

Uribe, H.A., B.W. Kennedy, S.W. Martin and DF. Kelton. 1995. Genetic parameters for
common health disorders of Holstein cows. J. Dairy Sci. 78:421-430.

VanRaden, P..M., B.L.Jensen, T.J.Lawlor and D.A.Funk. 1990. Prediction of transmitting
abilities for Holstein type traits. J.Dairy Sci. 73:191.

VanRanden P.M.1991. Derivation, calculation, and use of national animal model
information. J. Dairy Sci. 74:2737-2746

156

Van Arendonk. 1991. Use of proﬁt equations to determine relative economic value of dairy
cattle herd life and production from ﬁeld data. J. Dairy Sci. 74:1101.

VanRaden, P.M., E.L. Jensen, T.J. Lowlier, and D.A.Funk. 1990. Prediction of transmitting
abilities for Holstein type traits. J.Dairy Sci.73:191.

VanRaden, P.M., G.R Wiggans, and C.A.Emst.199l. Expansion of projected lactation yield
to stabilize genetic variance, J. Dairy Sci.74z4344.

VanRaden, P.M., and E.J.H. Klaaskate.1993. Genetic evaluation of length of productive life
include 11 predicted longevity of live cows. J .Dairy Sci. 76:2758.

Vanraden RM. and GR Wiggans. 1995. Productive life evaluation : calculation, accuracy,
and economic value. J Dairy Sci.78:631.

VanRaden, P.M., and A.E Freeman. 1985. Potential gains from producing bulls with only
sires as parents. J. Dairy Sci. 68:1425.

VanRaden, P.M., and E.J.H. Klaskate. 1993. Genetic evaluation of length of productive
life including predicted longevity of live cows. J. Dairy Sci. 76:2758-2764.

Van Tassell, CP., and L.D. Van Vleck. 1991. Estimates of genetic selection differentials
and generation intervals for four paths of selection. J. Dairy Sci. 74:1078.

Van Vleck, L.D., R.A. Welstell., and J.C. Schneider. 1986. Genetic change in milk yield
estimated from simultaneous genetic evaluations of bulls and cows. J. Dairy Sci.
6912963.

Weigel. K.A., T. J. Lawlor, JR, P. M. VanRaden, and G. R. ngans. 1996. Use of
Linear Type and Production Data to Supplement Early Predicted Transmitting
Abilities for Productive Life. J. Dairy Sci. 81 :2040-2044

Wiggans,G.R. and GE. Shook.1987. A lactation measure of somatic cell count. J. Dairy
Sci.70 : 2666.

Welper, RD. and A.E. Freeman. 1992. Genetic parameters for yield traits of Holstein,
including lactose and somatic cell score. J.Dairy Sci. 75:1342

157

Weigel,D.J., B.G. Cassell, and RE. Pearson. 1995. Relative genetic merit and

effectiveness of selection of young sires for artiﬁcial insemination. J. Dairy Sci.
78:2481

Weller, J.I. and D. Gianola. 1989. Models for genetic analysis of dystocia and calf
mortality. J. Dairy Sci. 7222633.

Westell R.A.; L.D.Van-Vleck.l987. Simultaneous genetic evaluation of sires and cows
for a large population of dairy cattle. J. Dairy Sci. 70: 1006-17

Westell, R. A., R.L. Quaas, and L. D. Van Vleck.1988. Genetic groups in animal model.
J. Dairy Sci. 71:1310.

Wiggans, G.R., P.M. VanRaden, and R.L. Powell. 1992. A method for combining United
States and Canadian bull evaluations. J. Dairy Sci. 75:2834-2839.

Wiggans, GR and M.E.Goddard.l997. A computationally feasible test day model for
genetic evaluation of yield traits in the United States. J .Dairy Sci. 80: 1795-1800.

Woolliams J.A. and Smith C. 1988. The value of indicator traits in the genetic
improvement of dairy cattle. Anim. Prod. 46:333-345

Woolliams, J.A. 1989. The value of cloning in MOET nucleus breeding schemes for
dairy cattle. Anim. Prod. 48:31- 35.

Woolliams, J.A.and Wilmut I. 1989. Embryo manipulation in cattle breeding and
production. Anim. Prod. 48:330.

Young, C. W. and J. G. Leece. 1960. Genetic and phenotypic relationships between
clinical mastitis, laboratory criteria, and udder height. J. Dairy Sci. 43:54.

Zhang, W. C., J.C.M. Dekkers, G. Banos, and EB. Burnside. 1994. Adjustment factors and
genetic evaluation for somatic cell score and relationships with other traits of
Canadian

158

"I11111111111111111111