Estimation of additive and non-additive effects on milk production traits in Iranian Holstein crossbred population

Direct and interactive effect (individual and maternal heterosis) was estimated using data from rotational crosses between Holsteins with Iranian native breeds. Traits of interest were milk yield, fat yield, fat percent and milk days. Complete data were available on 155 240 animals from 1991 through 2003. Direct additive’s effect, individual heterosis, maternal heterosis and recombination (interactions between presences of Holstein gene in two parents) effects were estimated by multiple regression method in SAS 8.2 with mixed models procedure. The least squares means of milk yield, fat yield, fat percent and milk days were 2 722.68±1 541.12 kg, 122.97±47.40 kg, 3.97±0.73 percent and 260.10±89.51 days respectively. All direct and indirect genetic effects are significant in milk and milk days traits (P<0.05). Individual and maternal heterosis and recombination effect are not significant on fat yield and fat percent traits. The individual and recombination effect were negative effect on milk yield. The result suggested that the Holstein is a favourable breed for crossbreeding program in developing country as Iran.


Introduction
With the increasing population in worldwide and need to increase milk production, the introduction of high-yielding breeds (i.e.Holstein-Friesian) plays an important role in protein needs supplying.In developing countries, the performance of these breeds is often negatively affected due to genotype-environment interactions (Bondoc et al. 1989, Rege 1991, Smith 1988, Ojango et al. 2005).One of the best methods for solving this problem could be crossbreeding.Crossbreeding as a mating system optimizes the additive genetic and nonadditive (heterotic) breed effects of Bos taurus and Bos indicus cattle in sustainable breeding systems (Gregory & Cundiff 1980).The design of optimal program and genetic manipulation would be more precise if the proportion of the genetic variance that is attributable to additive and non-additive gene action were known.In addition, knowledge of the maternal contribution to the performance of the offspring is necessary in planning sound breeding programs (Cassady et al. 2002, Solnker 1993, Vostry et al. 2008).The Iran dairy industry was based on approximately 1.4 million lactating cows in 2008.Breed composition of the herds was 65 % Holstein-Friesian, 27 % crossbred and 7 % native cattle (Naji 1996).Despite the high population of crossbred's cattle, a little estimates of breed and heterosis effects on productive traits have been reported, Rekui (2000) carried out first estimates of heterosis effects on the Isfahan crossbred cattle and reported significant effect on production traits.The objectives of this study were to estimate direct and maternal heterosis effects on production traits.

Materials and methods
A total of 155 240 animal records were collected from 1991 through 2003 from Holstein crossbred population by animal breeding centre of agricultural ministry, Iran.Data includes all province herds.For each cow milk yield (kg), fat yield (kg) fat percent and milk days traits information and breed composition were known.Editing data included checks on breed code of sires and progeny.Pedigree data were iteratively retrieved in seven loops for all cows with records and for their parents no additional pedigree was after this time.Ninety-five percent of the pedigree was found after three loops; give a total of 12 445 animals in model.Nine genetic groups were defined according to percentage of imported breed genes at intervals 12.5 %.

Statistical analyses
Records were analysed with following model that proposed by Van Der Werf & De Boer (1989): where µ is population mean, L i is lactation number i=1…8, H j and Y Sk are fixed environmental effects of herd and year season with j=1…295 and k=1…4, g l is fraction of Holstein gene in crossbred progeny with l=1…8 that equal to [(P s +P d )/2] (P s and P d are imported genes percent in two parent), H etm is hetrosis percent in progeny is equal to degree of heterozygosity of animal, R ecn is interactions between presence of imported gene in two parents, M heto is maternal hetrosis, a p is the additive genetic effect of cow making record; and e ijklmnop is a residual effect.Regression analysis was used here to efficiently estimate genetic effects given the mating systems employed in this study.Procedures for mixed models (PROC MIXED) were used to estimate breed direct, individual, maternal heterosis and breed-combination effects.

Results
The overall least squares means for milk yield, fat yield, fat percent and milk-days traits were 2 722.68±1 541.12 kg, 122.97±47.40kg, 3.97±0.73percent and 260.10±89.51days respectively which were in agreement with the means reported by Rekui (2000) and Naji (1996) in Holstein Friesian × Golpaigani cows.But, means slightly higher than those in the present study.These variations are to be expected as probable consequences of feeding and management changes, as well as changes in the genetic composition of the herd over the study.The linear regression approach was used to estimate of additive and non-additive effects to different traits.This statistical method will be able to separating component parts of performance and can be used to predict performance of various crossbred animals.Estimation of additive and non-additive effects on traits is given in Table 1.The effects are presented with their SE and P-value obtained directly from the regression models.The direct additive's effect on milk yield, fat percent and milk days traits were significant (P<0.05)but in fat yield trait is no significant (P>0.05).The value of this effect only on fat percent traits was negative.This situation arises due to the fact that Holstein has low fat percent compared with the native breeds.This result is not supported by the previous work of Rekui (2000) who worked with crossbred Holstein-Golpaigani cows and reported that direct effect on milk yield, fat yield and fat percent were 20.86, 0.712 kg and 0.041 percent receptivity.Significant regression coefficients were estimated for the effect of individual heterosis on milk yield and milk-days traits (P<0.001).The individual heterosis estimation on all traits, expect milk days trait, are negative.The maternal heterosis effects are significant on milk yield and milk day's traits but unlike Individual heterosis are positive in milk yield and fat percent traits and negative for other.Recombination effect is the breakup of epistatic effects during meiosis to form non parental in the locus combinations of alleles in gametes of crossbred parents (Dickerson 1973) or in the other hand it originate either through dominance effects, from interactions between native breeds and Holstein genes within loci, or epistatic effects from interactions between loci (Van Der Werf & De Boer 1989).This effect had same trend with maternal heterosis in significantly but in milk yield and fat percent traits were negative.Positive heterosis on milk yield were reported by MacDowell & McDaniel (1968), Beckett & Ludwick (1979), Katpatal (1977), Van Der Werf & De Boer (1989), Rincon et al. (1982), Ahlborn-Breier & Hohenboken (1991), Kress et al. (1996), Rekui (2000), Vanraden & Sanders (2003) but Soldatov & Dutsheev (1991) and Wang et al. (1992) reported negative heterosis.Beckett & Ludwick (1979), Rincon et al. (1982), Naji (1996) and Rekui (2000) reported positive heterosis on fat yield and MacDowell & McDaniel (1968) and Van Der Werf (1989) reported negative heterosis on fat yield.However Beckett & Ludwick (1979), Rincon et al. (1982) and Ahlborn-Breier & Hohenboken (1991) reported negative heterosis on fat percent trait and Rekui (2000) reported positive heterosis.
Positive maternal heterosis on milk yield and fat yield traits reported by Kress et al. (1996) and Rekui (2000) but Ahlborn-Breier & Hohenboken (1991) and Rekui (2000) reported negative maternal heterosis on fat percent.Contradictory results in papers may be due to difference of native and imported breeds used in crossbreeding program and number and breed composition of animal in analysis procedure.

Discussion
Results of this study are consistent with the results of further studies because most studies on the Holstein crossbred populations have reported positive heterosis while in this study for Holstein crossbreeds has been estimated negative heterosis in many traits.Recombination effect that shown interaction between imported and native genes in cross animal was same trend with individual heterosis.Totally, results of this study showed that entrance of new Bos taurus breeds in Iranian native cattle population has increased production economic traits, although this effect on the fat yield trait was not significant.
Crossbreeding is yet another tool in the genetic improvement and like anything else can be very profitable if understood and used correctly.