Methods of Selection

The net economic value of an animal depends on several traits, it is therefore, essential to estimate the total breeding worth of an animal based on several traits. Selectio based on more that one trait is known as METHOD OF SELECTION.

üIn practice, the breeding merit if animal is determined on several traits simultaneously and not by single trait, to improve the overall economic value based on simultaneous selection for several characters, also known as multi trait selection.

Tandem Selection: Selection for a particular trait is undertaken and when sufficient response is achieved, selection for other trait is undertaken.

• Less efficient tha other two methods of selection

• The average genetic improvement per generation in each of n traits would be only 1/n times.

• Requires more time for improvement because, if selection is done for one trait, till that time other trait has to wait.

• Efficiency also dependends most importantly on genetic correlation of traits.

Independent culling levels (ICL): A minimum standard for each trait is fixed and every animal to be selected must met this minimum level for each of the trait under consideration.

• Method is superior to tandem selection because selection is practiced for more than one traits.

• It allows to cul the animals at early stages

• No compensation of other traits which are either superior or inferior

• Difficult to fix culling level for any particular trait

• More emphasis are given on traits which are expressed in early life

• Method is useful for selection of cattle used for show purpose

Selection Index orTotal Score Method: Itdiffers from ICL, culling levels are flexible and each trait is weighted by score, also superiority in some traits can be make up by deficiency in others. Separate value for each of the traits is determined and these values are added to arrive at a total score for all the traits of an individual. The animals with the highest total score are selected to be used as parents of next generation.

The amount of weight given to each trait depends upon:

1. Relative economic value of trait

2. Heritability of the traits, and

3. Genetic and phenotypic correlation between traits.

If the characters are uncorrelated, each character is weighted by the product of its relative economic value and its heritability.

Selection index method is more efficient than the tandem as well as independent culling level methods.

Construction of Selection Index

• The index is the best linear prediction of an individual’s breeding value.

• Basically, the statistical principle of constructing selection indices is the same as that of fitting of a multiple regression equation for predicting a dependent variable from two or more known independent variables.

• Here it takes the form of a multiple regression of breeding values on all the sources of information

Let us consider the phenotypic value of i^th character (P_i) which is made up of two additive parts:

P_i = G_i + E_i

Let us assume that the genotypic value (G_i) is composed entirely of additive effects of genes and therefore G_i and E_i are uncorrelated.

Further, let a linear function of G_is be:

which defines the total genotypic economic score of an individual, a_i being the relative economic weight given to the i^th character. H is also referred to as the net merit of the individual.

However, we do not know the G_i ’s and as such H can not be used as a criterion of selection. Hence selection has to be based on some function of the observed values of various traits. It is simplest to have a linear function I given by:

where b_i’s are unknown coefficients to be determined. The values of these b_i’s are to be such that the function I may best discriminate the individuals with greatest genotypic economic score (H). It means that the selection based on I should be as nearly good as based on H, had it been known.

Let us suppose that we want to select I = 1 to K traits where X₁, X₂,…..X_k are the phenotypic values of these traits of an individual. The index of an individual is then:

I = b₁X₁ + b₂ X₂ +………. + b_k X _k

in which b’s are the factors by which each trait is to be weighted. They are the partial regression coefficients of the individual’s breeding value on each trait. To solve for b’s, we are required to prepare phenotypic and genotypic matrix along with b and a vectors.