MODELING OF SELECTION INDEXES FOR THE CAT EGORIES OF CZECH PIED CATTLE

The use and construction of selection indexes for farm animals is link ed up with the work of many authors. The objective of the present work is model construction of selection indexes for heifers and dairy cows , and for young, and proved bulls, and determination of expected selec tion effects by employing all information available from central files to a maximum extent. Partial breeding values are combined in indexes. All information available at the given moment is used to estimate bre eding value. Two two-stage indexes are constructed - for breeding cows and for sires. The reliability of aggregate breeding value for a trai t is determined by combinations of partial sources on the selection in dex principle (Pribyl et al., 1988). The variance of aggregate genotyp e is 1,040,596 Kc, the variance of heifer index makes 210,650 Kc; this indicates the reliability of aggregate breeding value estimate r(2) = 0.20. Selection reliability for cows is 0.36, for young bulls 0.16 an d for fully proved bulls 0.78. Breeding values of the particular selec tion steps and their sum are shown separately. Young bulls selected fo r insemination have the aggregate breeding value of 260 Kc, proved bul ls 1,771 Kc and fathers of bulls 2,036 Kc. In the trait milk performan ce the average breeding value of young bulls makes 38.6 kg, of proved bulls 498.9 kg and of fathers of bulls 579.6 kg milk, which is in agre ement with data on the breeding situation for Czech Pied cattle. Predi cted genetic gain is 199 Kc, milk production making 74% out of this va lue. Genetic gain of milk performance is 57.55 kg. Comparison of annua l genetic gains shows that annual genetic gain expressed in Kc for sel ected proved bulls equals 1511/199 = 7.6 times higher than that of you ng bulls in test. This trait for fathers of bulls is 8.9 times higher. Differences in generation intervals make 6.2 years. It is evident aft er this adjustment that the annual genetic gain of proved bulls is on average 1.4 times higher, i.e. by 278.6 Ki, than that of one generatio n younger bulls in test used parallelly. The difference in milk perfor mance between young and proved bulls is the eight times higher value o f annual genetic gain; after adjustment to genetic trend it indicates 1.8 multiple of the difference between parallelly used proved bulls an d by one generation younger bulls, i.e. the difference of 103.5 kg. Af ter adjustment to genetic trend, the heifers on average exceed the cow s-mates half a generation older both by their aggregate breeding value and by the breeding value of milk performance. The share of the group s of traits in aggregate selection effect is as follows: milk performa nce : meat performance : other traits = 81.12 : 20.03 : -1.16. Employi ng the presence structure of milk recording and information system man y items of information can be used that will increase the accuracy of breeding value estimate, and also of selection, in the process of sele ction of the particular categories of cattle. Animal model applying al l relationships between the animals should be used in general for the separate traits to estimate breeding values, which can lead to substan tial improvement mainly in young animals.