First Results of a MAS Study in Dairy Cattle with Respect to Longevity (short communication)

A study where Casein loci were used to be markers for improving milk yield, fat content and protein yield in a marker assisted selection experiment focused on possible side effects of this process on longevity traits. Using the data of two samples of cows a Simulation on the data was carried out evaluating the selection differences in period of use, period of life and production of the whole life indirectly. The selection criteria were flexible using one of the milk production traits each as well as marker information. During the investigated period no negative effect of simulated selection using EBV for milk yield, fat content and protein yield alone or in combination with the Casein loci used as markers on longevity traits occurred in the population. There are no significant differences regarding the longevity traits caused by selection for Single milk trait based on pure EBV and MAS.


Introduction
The progressive development of the number of mapped quantitative trait loci (QTL) is combined with the question of their application in dairy cattle breeding in the form of marker assisted selection (MAS).The productivity is considered to be on the first rang of selection schemes and applying MAS in general is aimed at an improvement of Performances in economic important traits.Using MAS strictly in one direction improves the productivity of the next generation but may cause a loss of genetic variants.An exclusive selection of desired genotypic variants using markers may also lead to an increased number of unknown linked loci which might increase a disease susceptibility or decrease positive genetic variants with unknown function so far.Such effects might be detected very late and may be followed by considerable negative economic responses.A compensation of them would take quite a long time in dairy cattle breeding.For this reason the examination for possible side effects on traits which are 'non-selection traits' in the sense of production, e.g.fertility and longevity, caused by a MAS focusing on the improvement of production traits, is more than necessary.New studies dealing with the genetic basis of longevity underline the importance of these traits (e.g.ELO et al., 1999).Longevity is influenced by several other factors of importance (e.g.SHORT and LAWLOR, 1992;DEKKERS, 1993).It is also a mi.rror image of the management conditions.The relative breeding value for period of use is now involved into the comprehensive breeding value used in the German dairy cattle breeding system by the Computing center for animal breeding VIT Verden (1998).
A number of studies focusing on the theoretical expectations applying simple schemes of MAS have been carried out (e.g.SPELMAN and BOVENHUIS, 1998;XIE and XU, 1998).The implementation of MAS in general may be successful if and only if the QTL focused on is verified.Falsely identified QTL applied in MAS may cause genetic loss versus a breeding scheme without any knowledge of QTL (SPELMAN and VAN ARENDONK, 1997).But the consideration of the MAS effectiveness in general is not the topic of this paper.As it was mentioned above possible negative impacts of MAS on non-selection traits being of some importance have to be excluded before applying it.Now in this paper we concentrate on possible side effects on longevity.Acceptcd parameters for this trait complex could be the period of use (sum of whole milking days), period of life (period from birth to the end of life) and the whole milk production (kg).Our first results are based on a simulated selection experiment on real data.The multiple allelic casein loci (CN) ß-CN and K-CN on the bovine chromosome 6 were used as markers for milk yield and fat content in different scenarios of a simulated selection.The positive selection response and the effectiveness of the selection on these milk traits using MAS in comparison to selection response using EBV alone are considered to be subject of a further paper.

Material and Methods
Information on 454 German Black and White cows from the herd for preservation this breed (SR) and 974 Black Pied Dairy cows (SMR) containing at least 75% Holstein Friesian genes fulfilled the demands of analysing longevity and could be included into this first analysis.Tables 1 and 2 show the production level for the longevity traits used here.A different allele frequency in both groups of cows especially in K-CN and ß-CN is to be considered.As could be shown in earlier studies (e.g.PANICKE et al., 1996) the changed production level is reflected by an alteration of allelic frequency.The age at first calving was averagely 915 days, respectively.The birth period ranged from 1980 to 1990 and all cows with culling data tili March 1999 were involved.The programmes VCE and PEST (GROENEVELD, 1992(GROENEVELD, , 1993) ) were used for prediction of animal effects.The general population genetic model allowing for fixed environmental effects (herd, year and season of calving) and the relation between animals has been adapted.The animal effects were the basic individual values for analysing the selection effects on longevity traits (Tables 3 and 4).MAS focused on improving milk traits with respect to estimated breeding values of milk traits (EBV) was carried out including related effects of marker genotypes in a simulated selection on the data.The selection intensities (SI) were 50 and 25 %.The results of our former studies mentioned dealt with the estimation of effects of CN-loci on milk traits and results of segregation and linkage analyses were considered for the choice of selection criteria.Milk yield was chosen as a selection trait because of highly significant effects of ß-CN on this trait in favour of the ß-CN A2 allele (PANICKE et al., 1996).An evidence for genetic linkage of K-CN and QTL for fat content was given (FREYER et al., 1996).In this population the positive QTL allele is linked to the K-CN B allele.Linkage is a real prerequisite for applying MAS which has been fulfilled in this case.The reason for including protein yield was the fact that a high protein yield plays an indicator role for stability.A 'suggestive linkage' of ß-CN and QTL for protein yield was given by own results using the same material described above.The selection criteria may be described as follows: (1) milk yield EBV (Mkg) (2) b_MAS combines Mkg EBV and effects of ß-CN on milk yield (3) fat content EBV (F%) (4) k_MAS combines F% EBV and effects of K-CN genotypes on fat content (5) protein yield EBV (Pkg), SMR only (6) bP_MAS combines Pkg EBV and effects of ß-CN genotypes on protein yield, SMR only The basis for evaluating the effects of the selection criteria 1 to 6 on the longevity traits is the comparison of the means and Standard deviations between the selected groups 1 to 6 following the selection criteria.The selection differences for longevity traits are therefore 'indirectly estimated selection differences'.

Results and Discussion
As described earlier in the introduction part this paper is dealing with the results of a Simulation on the data selecting cows for milk traits both alone on the basis of their EBV and using markers with respect to longevity traits only.The selection response on longevity traits is gained indirectly.
The results of simulated selection on the data both for SR and SMR show the same outline as is presented here for period of use (Table 5).Looking at the results from involving the markers into the selection criterion for milk production traits this is likely due to the same direction of the genotypic effects of ß-CN on milk yield, protein yield on one hand and K-CN on milk fat content on the other in both samples.Because of that similarity also for period of life and the whole productivity which are reflecting the period of use those traits are not shown as extendedly.We concentrate on the results of SMR because of containing more data and representing a cow stock reflecting the present conditions regarding increased impact of Holstein Friesian cows on the dairy cow population in the East of Germany (Table 6).Only small differences are gained.The only fact what should be mentioned is that the results suggest, a selection for fat content whether the marker is used or not, causes a slightly higher selection response on the longevity traits.Only a small superiority of e.g.11 to 14 days of use might be noticed, compared to a selection focused on milk yield.Selection on milk yield caused small positive deviations from the population mean in all longevity traits, both in 50 % and 25 % of the positive selected cows.This is shown by indirectly estimated selection differences for example in period of use (Fig.).The selection for fat content equally shows slightly positive effects on days of use.The protein yield used as a selection criterion both alone and combined with the marker Information affects the period of use moderately as well, whereas the whole milk yield is higher than using both other selection traits.The small effects of all selection criteria on period of life may be ignored (Table 6).We consider that no negative effects on any longevity trait were detected using the CN loci in our selection experiment focused on improving milk Performances.The present selection strategy in German Holstein involving the relative breeding value of longevity into the comprehensive breeding value (VIT, 1998) might be a guarantee for allowing for this important trait.So it should be no reason to assume that MAS applying e.g. the CN loci as markers focused on improving milk production traits would impair longevity.

Short Conclusion from first Results
During the investigated period no negative effect of a simulated selection using EBV for milk yield, fat content and protein yield alone or in combination with the CN loci, used as markers for milk traits, on longevity traits occurred in the population.There are no differences to be considered in this stage of the analysis caused by selection on pure EBV and MAS within a trait.Further examinations are necessary.
Fig.: Selection differences in days of use estimated indirectly whereas selection was carried out in favour of milk production traits (SMR) (Indirekte Selektionsdifferenzen für das Merkmal Nutzungstage nach Selektion auf Milchleistungsmerkmale (SMR))

Table 2
Milk production level of cows involved given in average (Standard deviation) of the first lactation (Milchletstungsniveau der einbezogenen Kühe als Mittelwert (Standardabweichungen) innerhalb der Stichproben SMR und SR bezüglich Langlebigkeit)

Table 3 Averages
and Standard deviations SD of animal effects in period of use, period of life and life Performance (milk yield kg) in SR cattle (genotypes n > 10 involved) (Mittelwerte und Standardabweichungen SD der Tiereffekte für Nutzungsdauer, Lebensdauer und Lebensleistung für die Stichprobe SR (über 10 Tiere hier gezeigt))