Estimates of genetic parameters for direct and maternal effects on birth weight of local sheep in United Arab Emirates

Small Ruminant Research 39 (2001) 219±224

Estimates of genetic parameters for direct and maternal effects on birth weight of local sheep in United Arab Emirates S.A. Al-Shorepy* Department of Animal Production, Faculty of Agricultural Science, United Arab Emirates University, P.O. Box 17555, Al Ain, United Arab Emirates Accepted 19 September 2000

Abstract Genetic parameters for birth weight were estimated for two sets of purebred and crossbred lambs in United Arab Emirates, using animal model methods. Data were analyzed by restricted maximum likelihood (REML). Five different animal models were ®tted. Model 1 considered the animal as the only random effect. Models 2 and 3 included in addition to the additive direct effect of the animal, the additive maternal and the permanent maternal environmental effects, respectively. Model 4 ®tted both the additive maternal and permanent environmental effects. Model 5 was the same as model 4, except that a covariance between the direct and the maternal additive effects was included. Estimates of direct heritability were substantially higher when maternal effects were ignored. Introducing the additive maternal effect to model 2 reduced the ^2p for the purebred and crossbred lambs, respectively. Estimates of additive estimate of additive heritability by 28 and 14% of s direct and additive maternal heritabilities with model 4 were 0.10 and 0.33, and 0.45 and 0.10 for purebred and crossbred lambs, respectively. The correlation between direct and maternal genetic effects for the combined (purebred ‡ crossbred lambs) data set was large and negative. These results indicate that in addition to additive direct effect, additive maternal effect for birth weight was important. # 2001 Elsevier Science B.V. All rights reserved. Keywords: Birth weight; Direct effect; Maternal effects; Purebred lambs

1. Introduction Sheep numbers in United Arab Emirates (UAE) has been estimated at about 193,000 out of a total livestock population of over a million and a half (AOAD, 1997). Local purebred sheep in UAE are black with small ears, long face and coarse-wool. They are thintailed (with slightly thicker base), small in size, early maturing animals and unthrifty in appearance; females are unhorned. They are well adapted to the harsh

* Tel.: ‡971-37051255; fax: ‡971-3632384. E-mail address: [email protected] (S.A. Al-Shorepy).

environment and their meat is preferred by the local population (Alhadrami et al., 1997). Birth weight has received limited consideration in sheep breeding programs, but is a trait of potential economic importance through its effects on preweaning growth and hence, increases the economic success of producing slaughter animals. An intermediate optimum has been shown to exist for birth weight, with excessively large lambs liable to dystocia and excessively small lambs at risk of death from hypothermia, starvation, respiratory diseases, and other causes. Smith (1977) and Notter and Copenhaver (1980) found that lamb survival was maximized at birth weights of 5.2 and 5. 5 kg, respectively, but that means

0921-4488/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 4 4 8 8 ( 0 0 ) 0 0 2 0 6 - 6

220

S.A. Al-Shorepy / Small Ruminant Research 39 (2001) 219±224

for birth weights in the population analyzed were only 4.1 and 3.9 kg, respectively. Therefore, it appears likely that an increase in birth weight would be desirable in many populations and that underweight lambs are a more serious problem than overweight lambs. Studies of various sheep breeds have shown that both direct and maternal genetic in¯uences are of importance for lamb birth weight (Maria et al., 1993; Tosh and Kemp, 1994; NaÈsholm and Danell, 1996; Al-Shorepy and Notter, 1998). It is generally concluded that the covariance between direct and maternal genetic effects is negative (Robison, 1981; Meyer, 1993; Maria et al., 1993). However, positive relationships have also been found (NaÈsholm and Danell, 1996; Yazdi et al., 1997). Therefore, the aim of this investigation was to estimate genetic parameters for direct and maternal effects on birth weight of sheep obtained from ®tting four animal models, attempting to separate direct genetic, maternal genetic and permanent environmental maternal effects. In addition, the genetic correlation between direct and maternal effects was estimated. 2. Materials and methods Data in this study were the accumulated records over the years 1994±1999 obtained from the Faculty of Agricultural Experimental Station located in Al Ain, UAE. It included 318 lambs of the local purebred and 339 crossbred (1=4 Chios and 3=4 local purebred sheep) lamb. The ¯ock included about 80 ewes with four local purebred rams, which produced approximately 120 lambs yearly. All the animals in this ¯ock were kept under intensive husbandry, with three lambings per 2 years. Ewes were ®rst exposed to rams at about 12 months of age. During mating, ewes were divided into four groups which were randomly allotted to rams, except that parent off-spring and sib matings were avoided. 2.1. Statistical methods Variance components and genetic parameters were estimated for each data set and for all data sets using DFREML (Meyer, 1998). Birth weights were adjusted

for effects of sex, type of birth (single, twin, or triplet) and age of ewe (1, 2, 3 or more years) before analyses. The analytical model included ®xed effects of season and (for the combined data sets) breed. Five animal models were ®tted. Model 1 was a simple animal model with additive direct genetic effects as the only random effect. Model 2 ®tted in addition, the effect of dam as an uncorrected random effect. Model 3 considered maternal genetic effects as the second random effect and allowed for covariance between the direct and maternal genetic effects. To test the signi®cance of random effects and to identify the most appropriate model, likelihood ratio tests were used after deleting each random effect (excluding residual) from the model. The general representation of the full general model (model 4) used is as follows y ˆ Xb ‡ Za a ‡ Zm m ‡ Zc c ‡ e; where y is a N  1 vector of records, b denotes the ®xed effects in the model with association matrix X, a is the vector of direct genetic effects with the association matrix Za, m is the vector of maternal genetic effects with the association matrix Zm, c is the vector of permanent maternal environmental effects with the association matrix Zc, and e denotes the vector of residual (temporary environmental) effects. Additive direct and maternal effects were assumed to be normally distributed with mean 0 and variance As2a and As2m , respectively, where A is the additive numerator relationship matrix and s2a and s2m are additive direct and maternal variances, respectively. Permanent environmental effects of the dam and residual effects were assumed to be normally distributed with mean 0 and variances I d s2pe and I n s2e , respectively, where Id and In are identity matrices with orders equal to the number of dams and individual records, respectively and s2pe and s2e are maternal permanent environmental and residual variances, respectively. Estimates of 2 ^ 2 ) and peradditive direct (^h ), additive maternal (m 2 b ) heritabilities were calcumanent environmental (pe lated as ratios of estimates of additive direct (^ s2a ), 2 additive maternal (^ sm )and permanent environmental maternal (^ s2pe ) variances, respectively to the phenotypic variance (^ s2p ). The direct-maternal correlation (^r am ) was computed as the ratio of the estimates of direct-maternal covariance (^ sam ) to the pro^2m . ^2a and s duct of the square roots of estimates of s

S.A. Al-Shorepy / Small Ruminant Research 39 (2001) 219±224

Maternal across-year repeatability for ewe performance, tm ˆ …14†h2 ‡ m2 ‡ c2 , was calculated. Heritability of the total additive genetic contribution to a maternally in¯uenced trait was calculated according to the following equation (Willham, 1972). ^h2 ˆ t

^2a s

‡

0:5^ s2m ‡ ^2p s

1:5^ sam

records animals sires dams progeny/sire progeny/dam

Purebred

Crossbred

Combined

318 398 19 110 16.7 2.9

339 425 16 94 21.2 3.6

657 821 22 204 29.9 3.2

2.91 a (0.88) 2.82 b (0.87) 3.12 a (0.84)

2.73 (0.81) 2.67 b (0.81) 2.82 a (0.80)

Adjusted to a twin-male basis. Means in the same row denoted by the different letters differ signi®cantly (P < 0:05). c Means in the same column denoted by the different letters differ signi®cantly (P < 0:05). b

1

s2p

s2p

3

Crossbred

Combined

0.42

0.75

0.59

0.38 0.38 ÿ8.25b

0.61 0.61 ÿ87.26b

0.55 0.55 ÿ92.21b 0.58

^ h ^2 m ^tm ^2 h t log L

0.09 0.35 0.37

0.48 0.16 0.28

0.31 0.20 0.28

0.55 0.20

0.69 ÿ84.1

0.65 ÿ84.8

s2p

0.42

0.73

0.57

0.24 0.29 0.35

0.52 0.15 0.28

0.40 0.15 0.25

0.38 1.62

0.61 ÿ84.2

0.55 ÿ86.5

0.42

0.72

0.57

0.10 0.33 0.02 0.38

0.45 0.10 0.08 0.29

0.32 0.17 0.03 0.28

0.55 0.20

0.69 ÿ83.9

0.65 ÿ84.75

0.42

0.75

0.59

2 ^ h m ^2 b2 pe ^tm 2 ^ ht log L

s2p

5

Purebred

0.72

s2p

4

Data set

0.42

2 ^ h b2 pe ^tm 2 ^ ht log L

2 ^ h ^2 m ^rdm b2 pe ^tm 2 ^ ht log L

Data set

Means and S.D. (kg)a Allb 2.44 b (0.58) 2.40 (0.57) Season 1c Season 2c 2.49 (0.60) a

Itema

2

Table 1 Characteristics of the data sets Ð means and standard deviation for lamb birth weight

No. No. No. No. No. No.

Model

2

Season of birth had a signi®cant effect on birth weight for the crossbred and combined data sets (P < 0:05) and approached signi®cance for the purebred lambs (P < 0:10). Lower birth weights were observed for lambs born in season 2 (from October to April) (Table 1). Reduced birth weights have been documented for lambs exposed to high environment temperature during gestation (Shelton, 1964; Alexander and Williams, 1971). Crossbred lambs were signi®cantly heavier than purebred lambs (P < 0:05). Variance component ratios for individual and combined data sets are shown in Table 2. All models included a random additive effect. Estimates of the direct additive heritability depend very much on the model used, with a range of 0.10±0.37 and 0.45±0.62 for purebred and crossbred lambs, respectively. Model

Item

Table 2 Variance component ratios and log likelihood for models for lamb birth weight (kg)

2 ^ h 2 ^ ht log L

3. Results and discussion

221

0.42 0.33 ÿ0.60 0.00 0.44 0.17 ÿ81.36

a

See text for de®nitions. Signi®cant (P < 0:05) change in likelihood relative to other models in the same column. b

1, which ignored maternal effects, resulted in high estimates of additive heritability in the three data sets. Crossbred lambs had higher heritability than other lambs. Heritability estimates were larger than those observed by Yazdi et al. (1997) for Baluchi lambs, whose estimates of h2 was 0.17. Fitting a similar model, Al-Shorepy and Notter (1998) reported additive heritability estimates for three data sets in

222

S.A. Al-Shorepy / Small Ruminant Research 39 (2001) 219±224

effect in all data sets. As a result, maternal permanent environmental effects were reduced, whereas additive maternal effects were of the same magnitude as those observed in models 2 and 3 for all data sets. Additive maternal effects were more important for purebred lambs, counting for 0.33 of phenotypic variance suggesting that the additive maternal in¯uence on birth weight is a real effect. A correspondingly lower value of h2 was also indicated. The maternal additive heritability for purebred lambs in this study largely agrees with the estimate presented by Burfening and Kress (1993) and NaÈsholm and Danell (1996). On the other hand, estimates obtained from crossbred lambs and combined data differed from those obtained from purebred lambs, and hence, additive effects were more important than maternal effects for these lambs. Means and phenotypic variance were larger in crossbred lambs. The ®t of model 4 did not improve the model compared with model 2 and 3 in all data sets, as shown by the values of log L. In Table 3, results from studies concerning direct and maternal genetic effects on birth weight are presented. Except for crossbred lambs, the additive heritabilities presented in Table 2 for model 4 largely agree with the estimates in Table 3. However, ®tting a similar model, Al-Shorepy and Notter (1998) found a large additive heritability (0.48) for crossbred lambs born in spring, which is in agreement with the estimate of crossbred lambs in this study.

crossbred lambs (Dorset, Rambouillet, and Finnish Landrace) of 0.11±0.51. Mavrogenis et al. (1980), working with Chios sheep obtained estimates of h2 to be 0.13. The high additive heritability estimates found in this study can be explained by high management level at the sheep breeding station, creating small environmental variations. Hence, all animals were raised under similar environmental, nutritional and management conditions. Introducing the additive maternal effect to model 2 reduced the estimate of additive heritability by 28 and 14% of s2p for the purebred and crossbred lambs, respectively. Judged by log L, ®tting additive maternal effects (model 2) resulted in a signi®cantly better ®t compared with model 1 in all data sets (P < 0:05). Replacing the additive maternal effects in model 2 with maternal permanent environmental effects in model 3 increased the h2 from 0.09±0.24 and 0.42± 0.50 in the purebred and crossbred lambs, respectively. Estimates of the maternal permanent environmental effect, coded as incidents that affect all records of progeny of the same ewes (possibly due to uterine capacity, feeding level at late gestation and maternal behavior of the ewe), explained 29 and 14% of the phenotypic variance in the purebred and crossbred lambs, respectively (model 3). In model 4, the genetic and environmental components of the effect of the dam (m2 ‡ pe2 ) are pulled apart and this resulted in reducing the additive genetic

Table 3 Reported estimates of genetic parametersa for direct and maternal effects on birth weigh in different breeds of sheep Breed and model

2 ^ h

^2 m

^r dm

^c2

Author

Romanove (Animal model) Finewool (Animal model)

0.04 0.07

0.22 0.30

ÿ0.99 0.11

0.10 0.00

Baluchi sheep (Animal model) Synthetic line with 50% Dorset, 25% Rambouillet and 25% Finnish Landrace (Animal model) Rambouillet, Targhee and Columbia (Maternal and paternal half-dibs, full-sibs, and offspring on dam and sire) Hampshire (Animal model) Dorset (Animal model) Romanove (Animal model) Afrino (Animal model) Present study (Animal model)

0.17 0.13 to 0.51

0.10 0.07 to 0.13

0.08 0.06 to 0.40

0.19 to 0.31

0.30 to 0.65

ÿ0.74 to ±0.18

Maria et al. (1993) NaÈsholm and Danell (1996) Yazdi et al. (1997) Al-Shorepy and Notter (1996) Burfening and Kress (1993)

0.39 0.12 0.07 0.22 0.42

0.22 0.31 0.13 0.09 0.33

ÿ0.56 ÿ0.35 ÿ0.13

0.37 0.27 0.32 0.12 0.00

Tosh and Kemp (1994) Tosh and Kemp (1994) Tosh and Kemp (1994) Snyman et al. (1995) Present study

0.17

ÿ0.60

^h2 : direct heritability; m ^ 2 : maternal heritability; ^rdm : direct maternal genetic correlation; ^c2 : permanent environmental variance as a proportion of phenotypic variance. a

S.A. Al-Shorepy / Small Ruminant Research 39 (2001) 219±224

Relatively few estimates of variance components for maternal effects on birth weight have been reported. Although, the additive maternal effects seem to have an important in¯uence on lamb weight in the three data sets, the genetic basis for these maternal effects could not be documented from these data sets. Notter and Hough (1997), suggested that partitioning maternal effects into additive and permanent environmental components requires large amounts of data with repeated records on individual ewes and the presence of related ewes in the data. Yet, even when these requirements are met, results are inconsistent (Al-Shorepy and Notter, 1998). Fitting the direct-maternal covariance in model 5 resulted in a negative estimate of the corresponding correlation and increased estimates of the direct and the maternal effects in all data sets. Unreasonable increases of direct additive and additive maternal variances were obtained for purebred and crossbred lambs and hence, the estimates were not presented for these two data sets. The small number of observations in these two data sets may have contributed to the in¯ation of direct and maternal estimates. The ®t of model 5 to the combined data set was signi®cantly (P < 0:05) better than of other models. The same conclusion was also drawn by Robinson (1996) in beef cattle data. Negative estimates for direct-maternal covariance for birth and early growth traits of cattle and sheep are many in the literature: Meyer (1993), and Diop and Van Vleck (1998) for cattle; and Maria et al. (1993) and Tosh and Kemp (1994) for sheep. A very important problem, addressed by Meyer (1992), is the dif®culty of statistically separating the direct and maternal component. This was shown to be true even for the data from speci®cally designed experiments (Al-Shorepy and Notter, 1996; Al-Shorepy and Notter, 1998). Except for model 5, estimates of total heritability h2t were larger than the estimates of additive direct heritability. In conclusion, this study showed that additive maternal effects are important for birth weight and need to be considered in any selection program. Acknowledgements This study was made possible in part through support provided by the of®ce of Assistant DVCAA for

223

research, United Arab Emirates University. The author would like to thank Dr. Ghaleb Alhadrami for his helpful suggestions during the review of this manuscript. References Alexander, G., Williams, D., 1971. Heat stress and the development of the conceptus in domestic sheep. J. Agric. Sci. Camb. 76, 53±72. Alhadrami, G.A., Nigm, A.A., Kholif, A.M., Abdalla, O.M., 1997. Effect of roughage to concentrate ratio on performance and carcass characteristics of local lambs in the UAE. Arab Gulf J. Sci. Res. 15, 137±148. Al-Shorepy, S.A., Notter, D.R., 1996. Genetic variation and covariation for ewe reproduction, lamb growth and lamb scrotal circumference in a fall-lambing sheep ¯ock. J. Anim. Sci. 74, 1490±1498. Al-Shorepy, S.A., Notter, D.R., 1998. Genetic parameters for lamb birth weight in spring and autumn lambing. J. Anim. Sci. 67, 322±327. Arab Organization for Agricultural Development, 1997. Yearbook of Agriculture Statistics, Vol. 16. AOAD, Khartoum, Sudan. Burfening, P.J., Kress, D.D., 1993. Direct and maternal effects on birth and weaning weight in sheep. Small Rumin. Res. 10, 153± 163. Diop, M., Van Vleck, L.D., 1998. Estimates of genetic parameters for growth traits of Gobra cattle. J. Anim. Sci. 66, 349±355. Maria, G.A., Boldman, K.G., Van Vleck, L.D., 1993. Estimates of variances due to direct and maternal effects for growth traits of Romanov sheep. J. Anim. Sci. 71, 845±849. Mavrogenis, A.P., Louca, A., Robison, O.W., 1980. Estimates of genetic parameters for pre-weaning and post-weaning growth traits of Chios lambs. Anim. Prod. 30, 271±276. Meyer, K., 1992. Variance components due to direct and maternal effect for growth traits in Australian beef cattle. Livest. Prod. Sci. 31, 179. Meyer, K., 1993. Covariance matrices for growth traits of Australian Polled Hereford cattle. Anim. Prod. 57, 37±45. Meyer, K., 1998. DFREML User Notes. University of New England, Armidale, Australia. È ., 1996. Genetic relationships of lamb NaÈsholm, A., Danell, O weight, maternal ability, and mature ewe weight in Swedish ®newool sheep. J. Anim. Sci. 74, 329±339. Notter, D.R., Copenhaver, J.S., 1980. Performance of Finnish Landrace crossbred ewes under accelerated lambing. II. Lamb growth and survival. J. Anim. Sci. 51, 1043±1050. Notter, D.R., Hough, J.D., 1997. Genetic parameter estimates for growth and ¯eece characteristics in Targhee sheep. J. Anim. Sci. 75, 1729±1737. Robinson, D.L., 1996. Estimation and interpretation of direct and maternal genetic parameters for weights of Australian Angus cattle. Livest. Prod. Sci. 45, 1±11. Robison, O.W., 1981. The in¯uence of maternal effects on the ef®ciency of selection: a review. Livest. Prod. Sci. 8, 121±137.

224

S.A. Al-Shorepy / Small Ruminant Research 39 (2001) 219±224

Shelton, M., 1964. Relationship of birth weight to death losses and to certain production characters of fall-born lambs. J. Anim. Sci. 23, 355±359. Smith, G.M., 1977. Factors affecting birth weight, dystocia and preweaning survival in sheep. J. Anim. Sci. 44, 745± 753. Snyman, M.A., Erasmus, G.J., Wyk, J.B., Olivier, J.J., 1995. Direct and maternal (co)variance components and heritability estimates for body weight at different ages and ¯eece traits in Afrino sheep. Livest. Prod. Sci. 44, 229±235.

Tosh, J.J., Kemp, R.A., 1994. Estimation of variance components for lamb weights in three sheep populations. J. Anim. Sci. 72, 1184±1190. Willham, R.L., 1972. The role of maternal effects in animal breeding: III. Biometrical aspects of maternal effects in animal. J. Anim. Sci. 35, 1288±1293. Yazdi, M.H., EngstroÈm, G., NaÈsholm, A., Johansson, K., Jorjani, H., Liljedahl, L.E., 1997. Genetic parameters for lamb weight at different ages and wool production in Baluchi sheep. Anim. Prod. 65, 247±255.