ANIMAL BREEDING AND GENETICS
CHAPTER 1 – WHAT IS ANIMAL BREEDING?
Animal breeding means selecting animals with a good trait of interest, such as high milk
yield/speed/height, that will mate to become the parents of the next generation. The selected trait is
expected to be better in the next generation. Animal breeding is made to improve the genetic level of a
population and to maintain genetic diversity. There are important aspects to consider:
1. Animal breeding is successful if the trait you chose to select for is heritable, so it can be inherited
by the next generation. A trait is heritable if it depends on the genes of the animal (at least in part).
a. Trait = a characteristic of the phenotype of an individual (so anything you can measure on
an individual). The phenotype depends on the genes of the animal and the environment.
2. The animals you select for breeding need to have different genetic background to make selection
possible (to make sure the next generation improved)
3. Animal breeding works at population level (not automatically at individual level), because the
success of breeding can only be judged by looking at how the average level of the trait you selected
has changed from one generation to the next in the entire population.
In animal breeding, humans decide which animals can reproduce so there is artificial selection. However,
natural selection still plays a role.
- Natural selection = process by which animals that adapt better to the environment have a higher
chance to survive and produce more offspring, so the next generation will be more adapted than
the current one. Even though we select the animals that are more suitable for our goals, natural
selection is still working. Sometimes it even works in the opposite direction as artificial selection,
e.g. in cows a high milk production (selected by humans) often results in less ability to get pregnant.
Animal breeding in context
Domestication = the process of conversion of wild animals to domestic use. First domesticated species is
the dog (12000 years ago) some wolves started to get close to humans and eat their wastes. In the case
of dogs, the relationship that built between humans and dogs was not forced, but domestication of other
animals may be forced by humans, e.g. animals were captured and controlled by humans. Only animals that
were not aggressive were able to adapt to these conditions, so there was a (natural) selection on
temperament. Prerequisites for successful domestication are:
- Animals need to be able to adapt to the feed given by humans.
- Animals must be able to survive and reproduce in captivity. Animals that need a very large
territory are not suitable for domestication.
- Animals need to be calm; “flighty” animals are not suitable because they will try to escape.
- Animals need to recognize humans as their superior, so they must have a flexible social hierarchy.
Animal breeding/selective breeding originated in the 1700s records were kept about the performance of
animals to select the best ones for reproduction. The first herdbook (pedigree book) was for thoroughbred
horses in England in 1791. Then herdbooks were established for other horse breeds, cattle, pigs, dogs etc.
With herdbooks, breeds were established.
- Breed = group of animals of a certain species that (thanks to generations of selective breeding) has
become uniform in performance and appearance.
With animal breeding, we achieved many things such as different breeds of dogs, an increased production
from farm animals, better reproduction of the best animals etc. thanks to different advanced techniques.
,So there are many improvements from selective breeding, but also some negative aspects, e.g. improving
one trait in the animal may cause a worse performance in another trait at the same time, e.g. health
problems, fertility problems.
CHAPTER 2 – DEFINING THE BREEDING GOAL AND SETTING UP A BREEDING PROGRAM
Animal breeding and setting up a breeding program involves thinking about the system, the purpose of the
animal, deciding the breeding goal, and then setting up the breeding program to achieve the breeding goal.
Setting up a breeding program involves 7 steps:
1. Describe the production system First you must identify what is your production system, such as
its location (environmental conditions), what is the purpose of the animal (production, sport,
companion, working such as transport/guide dogs/draught, saving) and what animal species and
breed you want to raise in that production system.
2. Define the breeding goal
3. Collection of information/data recording
4. Determine the selection criteria and estimate the breeding
value
5. Selection and mating
6. Dissemination
7. Evaluation of breeding program
Define the breeding goal (step 2)
The breeding goal defines which traits have to be improved, in which direction (increase/decrease) and
the importance of each trait in the breeding program. You also need to define the direction in which you
want to change the population, so if you want to e.g. increase milk production/fertility/longevity etc.
Characteristics of breeding goal:
- It often includes more than one trait to be improved/changed at the same time.
- It should summarize all traits in a single criterion: H=breeding goal H=v1A1 + v2A2 + …, where
v=(economic) value of the trait, and A=true breeding value of the trait.
o The economic value is the expected profit of improving a trait with 1 unit (e.g. 1 kg of milk
more, 1 egg more etc.).
- It contains breeding values. Breeding values are a measure of the genetic merit of an animal for a
certain trait, and give an indication of what could be the performance of the offspring of that
animal. E.g. a bull with a breeding value of +1000kg of milk is expected to have daughters that will
produce +500 kg more of milk on average, because each parent transmits half of their genes.
- It contains (economic) values to weight the traits, i.e. to determine their (economic) importance.
- Breeding goals should aim at the future. To reach the breeding goal, some years may be necessary
so you have to take into account what could be the desired characteristics of animals in the future,
how consumer desires might change, how legislations might change.
- Breeding goals might contain economic and non-economic weights. Examples of economically
important traits are milk yield, meat/egg production. However, there may be concerns e.g. about
animal welfare, so your breeding goal can contain non-economic weights/importance such as
improving disease resistance and health.
,There are two ways to define the breeding goal: 1) expressing the breeding goal as a sum of the breeding
values and their economic value; or 2) expressing the breeding goal based on non-economic values. These
are described below.
Breeding goals based on economic values
In this case, the economic values of the traits you want to select are based on an economic model of the
production system. There are two types of economic models used to derive economic values: a profit
equation or a bio-economic model.
A profit equation relates the profit of an animal or farm to the level of certain phenotypic traits (P) in the
breeding goal:
𝑃𝑟𝑜𝑓𝑖𝑡 = (𝑅1 – 𝐶1)𝑃1 + (𝑅2 – 𝐶2)𝑃2 + (𝑅3 – 𝐶3)𝑃3 + ⋯ (profit equation)
The (R-C) is the difference between the marginal revenues and marginal costs generated when the trait
increases of one unit, and P represents the phenotypic trait. The economic values of each trait can be
derived as the partial derivative of the profit equation, e.g. for trait 1:
d (Profit )
v 1= (economic value equation)
d (P 1)
Economic values indicate how much money would be earned when the trait is improved of 1 unit (e.g. 1
kg of milk more, 1kg of meat more). NB: often, production systems cannot be easily specified by a simple
profit equation, so a bio-economic model is used. A bio-economic model is a system of interrelated
equations.
Example:
the net merit index for milk production traits in the NL consists of lactose yield (kg), fat yield and protein
yield. The breeding goal thus is H = Vlactose*Alactose + Vfat*Afat + Vprotein*Aprotein (A=breeding value)
the values of Vlactose, Vfat and Vprotein are based on the profit generated by having 1 extra kg of
lactose, fat, and protein in milk. The profit function is: Profit = (Rlactose – Clactose)Plactose + (Rfat –
Cfat)Pfat + (Rprotein – Cprotein)Pprotein, Where Rlactose are the revenues from a kg of
lactose, Clactose are the costs of producing a kg of lactose and Plactose is the actual lactose
production/content of the milk. The economic values are the partial derivatives with respect to the trait,
e.g. for lactose it is Vlactose = d(Profit)/d(Plactose) = Rlactose – Clactose. Same for the other traits in the
breeding goal (H).
E.g. the revenue that the farmer receives from the dairy factory when selling milk is 0.54 €/kg lactose
produced. The main cost of producing lactose are the costs for feed, e.g. 0.26 €/kg lactose produced. So the
economic value for lactose is Vlactose = 0.54 – 0.26 = 0.28 €/kg lactose, which means that if lactose
production in the milk increases of 1kg, the farmer has 0.28 € increase in net profit. If we do the same for
protein and fat, the breeding goal equation becomes: H = 0.28*Alactose * 2.1*Afat + 4.1*Aprotein
Breeding goals based on non-economic values
Desired gains in certain cases, it is hard to determine the economic value of traits based on a profit
equation, for example for traits such as health and welfare. Also, in some cases using an economic breeding
goal might result in an undesirable change in a trait, e.g. increasing the milk production of cows often
results in reduced fertility, and you want to avoid it even if increasing milk production increases your profit.
,Therefore, you formulate a breeding goal based on desired gains, such as “maintaining/improving the
current level of fertility”.
Non-economic values the importance of traits in the breeding goal can also be weighed based on non-
economic values such as the biological efficiency or the carbon footprint, or with weighing factors that take
animal welfare into account instead of the economic value of the trait.
NB: even if the breeding goal is based on non-economic values or desired gains, the profit of the production
system is not ignored, because farmers still want to improve traits that will benefit their profit in the future.
The remaining steps of the breeding program:
3. Data recording/collection of information. To select animals, we need to collect info on these
animals to estimate the breeding values of individuals and to avoid mating relatives, e.g. their
phenotype level (e.g. how much milk produced). (chapter 3)
4. Determine selection criteria and estimate the breeding value. The breeding values of each trait of
the breeding goal must be estimated. (chapter 8)
5. Selection and mating. Selection = choosing which animals will reproduce based on their breeding
values; mating = pairing males and females that will reproduce. (chapter 9 and 10)
6. Dissemination. In many species, the animals that produce (production population) are different
from those that are bred (breeding population). Therefore, if genetic improvement is obtained in
breeding animals, it has to be disseminated into productive animals so that the genetic
improvement is also achieved in the animals that actually produce e.g. milk.
7. Evaluation of breeding program. The results of the program need to be monitored to see if the
objectives are achieved.
CHAPTER 3 – INTRODUCTION IN GENETIC AND STATISTICS
Genetics
Each cell of an animal has a nucleus that contains a pair of chromosomes, which contain DNA. During
meiosis, the pairs of chromosomes recombine and then split up so that each gamete has a single
chromosome of each pair. During fertilization, the single chromosomes in the oocyte and sperm cell
combine and form a pair again.
DNA includes several genes, regions that contain genetic information. Each gene can be found in different
forms called alleles. A locus is a position on a chromosome, e.g. a gene. Individuals carrying two copies of
the same allele of a gene are homozygotes; if they carry two different alleles, they are heterozygotes. These
combinations of alleles result in different phenotypes. Alleles can also be:
- dominant or recessive
- co-dominance: a heterozygote Bb shows the phenotype of both alleles
- overdominance: the heterozygote has a higher genotypic value than a homozygote (e.g. BB weighs
40kg, Bb weighs 42 e bb weighs 36).
- Incomplete dominance: the heterozygote is not exactly in the middle.
- Epistasis: occurs when there are interactions between multiple genes
Meiosis is the process in germ cells to produce gametes (sperm and oocytes). Recombination is a process
during meiosis that creates different combinations of genes on a chromosome because of crossing-overs.
,If we have a population of animals, we can estimate their genotype and the allele frequencies in that
population. Example: consider a monogenic trait with alleles Z and z; the possible genotypes are Z/Z, Z/z
and z/z. In a population of 630 animals, 375 animals have genotype Z/Z, 218 have Z/z, and 37 have z/z. The
genotype frequency of the three genotypes in the population is: 375/630 = 0.595; 218/630 = 0.346 and
37/630 = 0.059. The allele frequencies can be calculated: Z/Z animals have 2 Z alleles, Z/z animals have 1 Z
allele, and z/z do have 0 Z alleles. The frequency of the Z allele is: 0.595 + (0.346/2) = 0.768. The Z/z animals
have 1 z allele and the z/z animals have two z alleles. Therefore, the frequency of the z allele is: (0.346/2) +
0.059 = 0.232. The frequency of alleles is noted as p and q, where p usually indicates the frequency of the
dominant allele, and q the frequency of the recessive allele.
In a population under random mating, and in the absence of selection, migration, mutation and random
drift, the frequencies of alleles and genotypes have a special relationship. Hardy and Weinberg, discovered
that under those assumptions, the genotypes ZZ, Zz, and zz relate to each other as p2 , 2pq, and q2. The
frequency of the genotype Z/Z = p*p = p2, of the genotype z/z = q*q = q2, of the genotype Z/z = 2*p*q =
2pq. However, in most populations the Hardy-Weinberg equilibrium is not happening.
Statistics
Traits can be continuous or discrete. Continuous traits are e.g. height, because and animal can measure
1.1m or 1.112 etc. (they can be virtually any value). A discrete trait is e.g. the number of offspring born or
the presence/absence of a disease (no decimal numbers).
Mean the sum of the values of a trait in a population divided by the number of animals in the sample.
Variance σ2 = 1/n * sum of(xi – mean)2 (where xi is any value of x in the sample)
Standard deviation square root of σ2
Coefficient of variation it is the standard deviation divided by the mean: CV = σ/mean, e.g. a CV = 0.55
means that the size of the st. dev. Is 55% of the size of the mean.
Many animal traits follow a normal distribution (bell shaped), which is symmetrical around the mean.
When considering two traits, it is possible to determine their relationship and how they influence each
other. This relationship can be determined using the regression, correlation, or the covariance.
- Covariance the covariance between traits x and y is equal to: cov(x,y) = E(xy) – E(x)*E(y), where
E(x) is the mean of the observations for trait x, E(y) is same for y and E(xy) is the mean of the
products of the observations for traits x and y (so you multiply the observsations of x and y
together and then their average)
o E stands for Expectation
o Covariance is a measure of how much two random variables change together; it can be
positive i.e. the two variables show a similar behaviour (e.g. they both increase together),
or negative i.e. they have opposite behaviour (e.g. one increases and the other decreases).
- Correlation formula is r(x,y) = 𝜌𝑥𝑦 = 𝑐𝑜𝑣(𝑥,𝑦)/𝜎𝑥𝜎y. The correlation is a value between -1 and 1.
NB: correlation does not show cause-effect relationships, i.e. the level of y is not influenced by the
level of x or vice versa. It just shows how strongly two variables may be related to each other.
- Regression it is a measure of the relation between x and y expressed as b(y,x) = 𝛽𝑦𝑥 = 𝑐𝑜𝑣(𝑦,𝑥)/
𝜎2x (regression coefficient)
o b(x,y) (beta) is the change in the value of y when x is increased of 1 unit.
o The regression coefficient can be negative or positive based on the relationship between
the two variables.
,In chapter 3 (pag 50-51) there’s an appendix with rules to calculate covariance, correlation and regression
based on the type of variables.
CHAPTER 4 – COLLECTION OF INFORMATION (STEP 3 OF BREEDING PROGRAM)
The third step to set up a breeding program is collecting information about the traits that are involved in
the breeding goal, e.g. if breeding goal is increasing milk production, you must collect info on the
production of your animals.
Information is collected using pedigrees, by measuring phenotypes and by sequencing the genotype of
animals.
An essential information in animal breeding is the pedigree of animals pedigree = the set of known
parent-offspring relationships in a population (e.g. a family tree). Pedigrees can be used to determine the
genetic relationships between individuals of the same population/breed. The pedigree is fundamental to
determine the breeding value of animals, in combination with the information collected about the trait of
interest. It is important that:
1. the pedigree has a reliable identification system, so each animal has an identification number
2. the measurements on the traits of an animal (e.g. milk production) must be associated with the
correct identification number (so there should be no mistakes in matching the measurements and
the animal on which they were taken).
Collection of phenotypes
Qualitative vs. quantitative traits:
- Qualitative traits = traits that can be divided into a few different classes and that are easy to
record, e.g. the colour of the animal (there’s a limited number of colors so classifying animals is
easy). They are usually determined by a very limited number of genes. Qualitative traits are usually
discrete variables in statistics.
- Quantitative traits = harder to record because there are many classes or are continuous so many
different values are possible, e.g. weight of animals can be literally any number. These traits are
polygenic = determined by several genes.
Measurement errors:
Measurement errors determine how accurately the phenotypes have been measured.
- Systematic errors = usually caused by differences between animals, e.g. you want to measure a
certain trait on two groups of animals, but they might have a different diet, age or training level, so
you will get different results because those different factors influence the phenotype.
- Random errors = they occur without any systematic cause, e.g. when you measure the height of a
cow 10 times, you might get slightly different results because the cow might move. So these are
errors that occur while actually taking the measurement.
Accuracy of measurements:
- Repeatability = the extent to which measurements on the same animal and under similar
conditions correspond with each other (e.g. I measure something on the same animal 10 times,
how similar the results are?).
- Reproducibility = the relationship between measurements in different locations and/or by different
persons (e.g. two people measure something on the same animal, how similar the results are?).
, Frequency of taking measurements:
- How frequently you measure a certain phenotype depends on the trait, the cost of measuring it
and the opportunity to measure it. For instance, if a farmer owns a machine that can record milk
production of cows, it can be made often; if you need to collect information on the conformation of
animals you have to call an inspector, but it is expensive, so you will do it only a few times a year.
Measurements are made on the animal itself or on its relatives?
- It depends on the type of trait you are considering. For example, you can measure growth in both
sexes at all moments in life. Milk production can only be measured on females after they give birth,
so for example to get the breeding value for milk production of a bull you can only consider the
milk production of its daughters.
- There are three sources of information that can be used to obtain info on the phenotype (and
consequently on the breeding value) of animals:
o Parents information, e.g. their milk production, fertility, longevity to predict that of the
offspring
o Siblings or half-siblings information, e.g. their milk production, carcass traits, fertility,
longevity. This info can be used because animals share genes with their siblings, so you
could predict the performance of the animal based on that of siblings
o Progeny information, because if you have info on the phenotype of the progeny you can
make a prediction on the phenotype of the parent (because they share half of genes)
Value of indicator traits:
- Indicator traits are used when we want to record another trait, that is related to the first one, that
is difficult to measure or that will be expressed only late in life, e.g. longevity of an animal. For
example, if longevity in horses is directly proportional to the quality of their legs, then leg quality
can be measured to predict the longevity of horses. We can also select for good quality of leg to
improve longevity in an indirect way. In this case, leg quality is the indicator trait.
Collection of DNA information
Information about DNA in animal breeding are collected with parentage tests, DNA test, with processes to
identify gene locations on the genome, to determine biodiversity in a population or as a tool to select only
the best animals for breeding.
NB: samples to get DNA are usually taken on somatic cells where all cells have the same DNA. Oocytes or
sperm are not the preferred methods because, since there is recombination during meiosis, each sperm cell
might contain a slightly different DNA in the chromosomes.
Samples to collect on the animal:
- Blood samples = very accurate but take a lot of time if they have to be taken from many animals.
- Roots of hairs = the DNA is in the hair follicles. It is easy, cheap, and non-invasive and it is suitable if
DNA samples have to be taken from a large group of animals, but it is less accurate than blood.
- Mouth swabs = non-invasive but difficult because all animals need to be handled individually and it
takes time. So at this point taking a blood sample is better because it’s more accurate.
Genetic markers:
After taking the samples, you need to read the DNA. One could determine the entire genetic sequence of
DNA, but it’s to expensive to do it for all animals. So the alternative is to use genetic markers.
