Community-based breeding programs (CBBPs) have recently attracted global interest as genetic improvement strategies in low input systems (Lamuno et al., 2018; Haile et al., 2019). Community-based breeding program has been designed to ensure involvement of farmers' (target groups) in all steps of the breeding program (Mueller et al., 2015).
In Ethiopia, in 2012, the International Center for Agricultural Research in the Dry Areas (ICARDA), in partnership with the Southern Agricultural Research Institute (SARI), Areka Agricultural Research Centre (AARC) of Ethiopia adopted CBBP in Doyogena district to improve Doyogena sheep. Doyogena sheep was among the potential breeds of the country with better market preferences in the local market and Addis Ababa. The sheep has attractive morphological features with a great potential for twining and fattening.
The potential genetic improvement of traits of interest was largely dependent on its heritability and genetic relationship among the traits of economic importance upon which selection may be applied. Information on heritability is essential for planning efficient breeding programmes, and for prediction of response to selection (Falconer and Mackay, 1996). According to Berkana (2019), evaluation of any designed genetic improvement program is fundamental either to optimize the program if the designed improvement program is progressing towards the set goals or redesign other alternatives if it fails or deviates from the preset goals. Moreover, evaluation of genetic trend gives an indication of genetic direction of the breed as well as the rate of genetic improvement from the time of application of the breeding program (Mallick et al., 2016). However, genetic studies of productive traits in sheep in low input systems are scarce due to lack of recorded data (Aguirre et al., 2017).
The data generated under Doyogena sheep CBBP for the last six years has not been statistically analyzed and/or evaluated to unravel the actual progress achieved in this program. Therefore, the present study has been planned to evaluate this program with objectives of assessing the effect of non-genetic factors influencing growth traits, estimate inbreeding level, estimate the genetic parameters and trends for growth traits and generate information for the optimization of the ongoing CBBP.
Material and MethodsDescription of study Area
The study area (Doyogena sheep CBBP)was found in KembataTembaro Zone of southern Ethiopia at a distance of 258 km to the Southwest of Addis Ababa (national capital).The district is located between 7°20’ N latitude and 37°50’ E longitude. Altitude ranges from 1900 to 2800m.a.s.l.The district received an average rainfall of 1221 mm between 2013-2017 and smallest total annual rainfall is recorded in 2017.The area is characterized by mixed crop livestock production system.Breeding program description and animal management:
Sheep flocks were generally managed by CBBP members. Animals were identified by plastic ear tag. In each of the breeder cooperative, an enumerator was employed for routine animal identification, data recording and follow-up. Enumerators use herd book for data recording. Selection of breeding rams takes place on a programmed date, twice per year. Researcher identifying candidate breeding rams by estimating best linear unbiased prediction (BLUP) breeding values for selection criteria traits using the performance and pedigree data recorded by the enumerators. In the first stage, candidate ram’s pre-selection and ranking takes place based on WWT. In second stage breeding rams ranking carried out based on six months weight (6WT) estimated breeding values (EBVs). Top (10% of the candidate) breeding rams were retained for breeding to be used in the community flock while next best (positive EBVs) were sold for breeding purpose to other communities. on-selected males (negative EBVs )were either castrated or marketed to prevent un wanted mating (Haile et al., 2019).Selected best breeding rams usually serve not more than one year in the community flock. After one-year service, the breeding rams were sold to other area of the region. This is because the rams become big in size and aggressive which makes difficult.
The main feed sources for animals included Enset (E ventricosum) products of Amicho, corm), crop residue, improved forage/grass, crop aftermath, kitchen leftover, and purchased concentrates. Flocks graze with tethering in the small private land. Free veterinary service was provided for CBBP participant farmers by ICARDA, and SARI.Data Sets
The empirical data for the study were obtained from the ongoing 5 breeder cooperatives. The performance data along with pedigree information is being maintained in data recording book of individual breeder cooperatives. The data routinely collected by the enumerators were recorded at the time of event. The birth weight (BWT) was recorded within 24 hours of lambing; weaning weight (WWT) was taken from 85-95 days of age; and 6-month weight was taken between 185-195 days of age. Since both WWT and 6-months weights were recorded on fixed dates, these records were adjusted for fixed age of 90 and 180 days for WWT and 6-months weight, respectively.
Reliability and consistency of pedigree information were checked by pedigree viewer software (Kinghorn, 2015). Records with duplicated and bisexuality were removed. For this study, available breeding data from the entire breeder cooperative were compiled filtered, and cross checked for its consistency and in formativeness.
The WWT and 6-month body weight were adjusted at fixed age of 90 and 180 days.
The average daily body weight gains from birth to weaning age; weaning to six months of age; and birth to six months of age has been estimated as below:
Average daily body weight (BW)gain up
to weaning age (gm) =
Average daily BW gain from weaning
to 6 month age (gm) =
Average daily BW gain form birth to 6 month age (gm) =
Where: BWY=Birth weight, AWWT=Adjusted weaning weight at 90 days,
A6MWT = Adjusted 6-month weight at 180 days
Statistical analysis of growth performance data
Effects of non-genetic factors
Data used for analysis included birth weight, three-month weight, 6-month weight, average daily gain from birth to weaning, average daily gains from weaning to 6-month age, and average daily gain from birth to 6-month age. Before conducting the main analysis, preliminary data analysis; like homogeneity test and normality test were employed using PROC UNIVARIATE in SAS (SAS, 2009).Abnormal records were eliminated as outliers. Then, remained data were analyzed using the Generalized Linear Model (GLM) procedures of SAS.
The non-genetic factors used in the model included, parity was classified as 1, 2, 3, 4, 5, 6 and ≥7 and because of small number of observations all parities above 7 were merged together under parity ≥7, site (Ancha, Hawora, Serera, Begedamo and Murasa).Similarly, birth type was classified as 1, 2, and ≥3 and because of small number of observations; litter size above triplet were included and considered as ≥3. The fixed effect of lamb birth season was classified in to three classes as main rainy season (June to September) and small shower falls (February to May) and dry season (October, November, December and January), year of birth (2013 to 2018) and sex (male and female). Tukey-Kramer test was used to separate least squares means with more than two levels. Fixed effects which were significant (P<0.05) were fitted in to the model to estimate the genetic parameters.
Model for analysis of variance for BWT, WWT, 6WT, ADG0-3, ADG3-6, andADG0-6Yijklmn= μ + Pi + Sj + btk + yrl +Sem+Sxn+ eijklmn Where: Yijklmn = growth trait for each animal μ = overall mean, pi = ith parity (i=7; 1, 2, 3, 4, 5, 6, ≥7) Sj = jth site (j= 5; Ancha, Hawora, Serera, Murasa and Begedamo) btk = kth birth> type (k =3; single, twin, triplet and above) Yrl = lth year (l=6; 2013 -2018) sem = mth season (m=3; main rainy season, Small shower falls, dry season) Sxn= nth sex (n =2; male, female) eijklmn = random error
Description of pedigree structure
The pedigree structure used for genetic parameter estimation and trend analysis is presented in Table 1.
Genetic parameter estimates
The variance components and resulting genetic parameters were estimated on a model fitting effects of parity, year of birth, season of birth, type of birth, sex, and site (cooperative) as fixed factors and growth performance traits (BWT,WWT, 6WT,ADG0-3, ADG3-6 & ADG0-6) as response variables using WOMBAT software (Meyer, 2012). Univariate animal model was fitted for all traits to estimate the genetic parameters. By including or excluding d maternal genetic or maternal environmental effects and with or without a covariance between animal and maternal genetic effect, the following six models were fitted for each trait:
Model 1: Y ꞊ Xb + Z1a + e
Model 2: Y ꞊ Xb + Z1a + Z3c + e
Model 3: Y ꞊ Xb + Z1a + Z2m + e with Cov(a,m) ꞊ 0
Model 4: Y ꞊ Xb + Z1a + Z2m + e with Cov(a,m) ꞊ Aσam
Model 5: Y ꞊ Xb + Z1a + Z2m + Z3c + e with Cov(a,m) ꞊ 0
Model 6: Y ꞊ Xb + Z1a + Z2m + Z3c + e with Cov(a,m) ꞊ Aσam
Where: y = vector of observed traits of animals; b, a, m and c are vectors of fixed effects, direct additive genetic effects, maternal additive genetic effects and maternal permanent environmental effects, respectively: X, Z1, Z2 and Z3 are incidence matrices relating fixed effects, direct additive genetic effects, maternal additive genetic effects and maternal permanent environmental effects, respectively; e=vector of residuals. Depending on the model, heritability was computed as: Direct heritability as, σ2a/ σ2p and maternal heritability; σ2m/ σ2p and, the direct-maternal covariance as proportion of phenotypic variance; (cam = σam/σ2p). The maternal environmental variance ratio was estimated by the maternal permanent environmental variance as a proportion of σ2c (c2 = σ2c /σ2p). The genetic correlation between direct and maternal genetic effects (ram) is estimated as a ratio of the estimates of the σam to the product of the square roots of the estimates of σ2a and σ2m .i.e. ram =
Total heritability (h2t) was calculated according to the following equation (Willham, 1972).
Where: σ2a = additive genetic variance, σm2= maternal genetic variance, σam = additive to maternal genetic covariance, σ2p = phenotypic variance..To determine the most appropriate model, likelihood ratio test (LRT) was used. The AI-REML algorithm was used in the subsequent analysis to check the convergence. Model comparison under Log L was considered to have a significant influence, when its inclusion caused a significant increase in Log L, compared to the model in which it was ignored. The Log L ratio=2 times maximum likelihood for full model minus maximum likelihood for reduced model and chi-square (χ 2) with degree of freedom equal to the difference between the two model being compared (Meyer, 2004). When Log L did not differ significantly (P>0.05), the model that has fewer parameters were selected as the most appropriate model. The genetic trends of the traits were estimated using regression of the average breeding value of the animals estimated from best model on the year of birth. Multivariate analysis in WOMBAT (Meyer, 2012) was applied to estimate genetic and phenotypic correlations based on the most appropriate model determined in the univariate analysis.
Results and DiscussionFixed effects
The overall least square mean of birth weight, weaning weight, and 6 month weight were 3.05±0.025, 14.8±2.49, and 22±0.22 kg respectively (Table 2).Single born lambs were heavier (P<0.05) than twin and ≥triplet in BWT and WWT. This effect may be attributed to lesser availability of uterine space (horns) among multiple births affecting prenatal nutrition development and also competition for dams' milk during pre-weaning period. Similar results were documented by Shigidafe et al., 2013 and Berhanu and Aynalem (2009). Season effect was significant for WWT (P<0.05) and not for the remaining growth traits (P>0.05), WWT was highest (P<0.05) for the lambs born during the small shower rainfall (13.95±0.45) than the lambs born during the dry and main rainy seasons. This indicated that small shower rainfall possibly provided optimum conditions for development of lambs. The effect of season is associated with difference in feed and disease situation (Berhanu and Aynalem, 2009).Male lambs were heavier than female lambs and this may be due to the influence of hormones in the two sexes. Year of birth was a significant source of variation for BWT, WWT and 6WT. and Horro sheep breeds under CBBP (Haile et al., 2020). The overall average daily gain (in grams) from birth to weaning age, weaning to 6month age and birth to 6 month age were 130.37±2.27, 80.59±3.62 and 106.18±1.7gm/day respectively(Table 2).Single born lambs grew faster than twin and ≥triplets. Similar to growth traits males lambs had higher average ADG0-3, ADG3-6 and ADG0-6 than females. The current result was higher than the previous reports on other Ethiopian sheep breeds (Duguma et al., 2002). Male lambs had higher average ADG3-6 than females (Table 2).
The result of ADG3-6 was higher than the report of Shigdafe et al. (2013).Shigdafe et al. (2013) reported lower ADG0-3 in traditional management system compared with the current result. This difference might be attributed to both the genetics as Doyogena sheep is a large in size and also the better management as practiced by member farmers who are economically dependent (to a large extent) on the sheep sale. The overall average daily gain (in grams) from birth to weaning age, weaning to 6 month age and birth to 6 month age were 130.37±2.27, 80.59±3.62 and 106.18±1.7gm/day respectively (Table 2). Single born lambs grew faster than twin and ≥triplets. Similar to growth traits males lambs had higher average ADG0-3, ADG3-6 and ADG0-6 than females. Male lambs had higher average ADG3-6 than females (Table 2). This difference might be attributed to both the genetics as Doyogena sheep is a large in size and also the better management as practiced by member farmers who are economically dependent on the sheep sale. The current result was higher than the values reported by Haile et al. (2014) for preliminary result of Menz and Horro sheep.
Coefficient of inbreeding shows increasing trend within the 6-year selection period (Figure 1).The coefficient of inbreeding was assumed to be zero until the year 2014; afterwards it increased with the selection years. At the time of this study (2018 selection year) coefficient of inbreeding was 0.30% with average annual inbreeding trend 0.08%. The proportions of inbred animals in the studied animal were 37. Average inbreeding level for the inbred animals was 18.4% of inbreeding was found. The most likely reason for this inbreeding increment could be selection of superior breeding rams without seeing their detail pedigree. Studies have shown that, inbreeding levels, higher than 10% could lead to inbreeding depression (Negussie et al., 2002). The inbreeding coefficients obtained for Doyogena sheep considerably in acceptable percentage, however, F is in an increasing trend, and thus consideration should be given during allocation of breeding rams. Negussie et al. (2002) reported, 20 years (1978-1997) inbreeding coefficient of 0.78% with annual trend of 0.07% for Horro sheep on the station management system and Gizaw et al. (2013), reported inbreeding coefficient for Menz sheep after 10 year selection was 1.7% with 0.17% increment per generation that is higher than present estimated inbreeding coefficient. When inbreeding was on the rise in the populations and in general, inbreeding leads to reduction in additive genetic variance and heritability (Falconer and Mackay, 1996). The traits declined with increase in inbreeding and caution should be taken in during selection.
Log L value of all the different models on the traits considered are presented in Table 3. The most suitable models are marked in bold font. The full model (Model 6) for BWT had the highest Log L value. Allowing for direct and maternal additive genetic covariance (Model 4) for WWT significantly increase Log L when compared with other models. The most appropriate models for average daily weight gain traits showed that maternal permanent environmental influences were important for those traits.
Based on the appropriate models, the estimates of direct heritability (h2a) for BWT, WWT, and 6WT were 0.33±0.06, 0.31±0.06, and 0.14±0.06, respectively (Table 4). Except the moderate heritability estimates for 6WT, which reflect less genetic variation among lambs at 6-month age, the estimates of direct heritability for BWT and WWT fall within the range of values reported in the high heritability value. The results showed that from an early age to a 6-month age, the proportion of genetic effects on variation is decreased as the proportion of environmental effects increased. This indicate, that the effects of unknown environmental effects also direct effects on growth traits as the lamb aged from birth (0.12), WWT (2.72) and 6WT (5.11) continuously Increased.
The present estimate of direct heritability for BWT (0.33±0.06) was; found in the range reported by Yacob (2008) of a direct heritability estimated for Afar sheep (0.1-0.38) and BHS sheep (0.2-0.58) using univariate analysis. A very high estimated h2a of (0.46) from the multi-trait animal model was estimated for Menz sheep by Gizaw et al. (2007), while Abegaz et al. (2002) estimated h2a of 0.20±0.05 for Horro sheep using same model and reported lower estimate than the current estimate for Doyogena sheep. The current result was higher than the estimate of Assan et al. (2002) for Sabi sheep (0.28),El Fadiliet al. (2000) for Moroccian Timahdit sheep (0.18),Gizaw et al. (2014b) for Menz sheep (0.019±0.036) using multi-trait individual animal model analysis and Haile et al. (2020) for Bonga (0.29 ± 0.047),Horro(0.16 ± 0.040) and Menz sheep (0.07 ± 0.027) by fitting univariate animal model. The estimate of h2a for WWT (0.31±0.06) was also found in the estimated range reported by Yacob (2008) for Afar sheep (0.11 - 0.37) and BHS sheep (0.00 - 0.29) but lower than the estimated for Menz sheep (0.46) by Gizaw et al. (2007).The estimate by Abegaz et al. (2002) for Horro sheep (0.16±0.05) and Gizaw et al. (2014b) for Menz sheep (0.19) was lower than the present estimates.
The estimate of direct heritability(h2a) for 6WT (0.14±0.06) was found in the range of direct heritability estimated for Afar sheep (0.11-0.37) and BHS (0-0.29), while, the report of Gizaw et al. (2014b) for Menz sheep 0.46 was much higher than the present estimate. Abigaz et al. (2002) estimated 0.18±0.05 of direct heritability for Horro sheep that was lower than the current estimate. From the genetic point of view, higher direct heritability estimates for BWT and WWT indicates that high variation within the breed and will be a greater opportunity for selection response during genetic improvement through selection for these traits. Moreover, WWT will be the best criterion for selection to increase pre weaning growth rate because selection on the basis of BWT which has the highest heritability could cause dystocia. However, the confounding effect of direct genetic and maternal genetic effect needs to be consideration.
The permanent maternal environmental effect (c2) for BWT was moderate in this study (0.20±0.09). This indicates the importance of maternal environment and care at the birth of lambs. The current estimates were similar to the findings of Gowane et al. (2010a) in Bharat Merino sheep (0.19 ± 0.02).For the 6WT trait, the maternal environmental effect is more important than maternal genetic effects. A similar finding was reported by Venkataramanan (2013) for Nilagiri and Sandyno Indian sheep breeds. The result suggested that, even if maternal effects tend to diminish with age, some adult traits will nevertheless contain this source of variation.
The current finding of BWT and WWT indicated that maternal heritability (h2m) is important variance components and the estimates were 0.24±0.12 and 0.60±0.07 respectively. Compared with other study, the BWT maternal heritability estimate was ranged in the estimate for BHS sheep (0.06-0.46) estimated by Yacob (2008).Abegaz (2002); Assan et al. (2002) and Yacob (2008) estimated h2m of 0.12±0.2, 0.24, and 0.02-0.21 for Horro sheep, Sabi sheep, and Afar sheep respectively and these all value are lower than the present estimate. However, higher estimated values of h2m were reported for Moroccoian Timahdit sheep (0.59) by ElFadiliet al. (2000) and Farafra sheep (0.40±0.001) by Mousa et al. (2013). The present estimate of maternal heritability (0.6±0.07) for WWT was higher than the above-mentioned sheep breeds estimated maternal heritability. High and negative additive-maternal genetic correlation estimates were observed (Table 5) for BWT (-0.61±0.15) and WWT (-0.81±0.11) traits. Similar results were summarized by Safari and Fogarty (2003) for a wide range of sheep breeds.
The correlation estimates between direct additive and maternal genetic effect (ram) for both the traits become negative means improvement in one will result in reduction of another. The result might be due to the structure of the data set used in the analysis i.e. the number of generations the animals were measured both directly and as dams were limited caused lack of large pedigree. It is essential to have a high proportion of dams and dams of dams with records (Safari et al., 2007).However, the data set for the present study were collected over only a period of 6 years, It could be lacking the optimum pedigree structure for accurate and reliable estimates of direct-maternal covariance components (Maniatisand Pollott, 2003).
σ2a = direct additive genetic variance; σ2c= maternal permanent environmental variance; σ2m= maternal additive genetic variance; σam= additive and maternal additive genetic covariance, σ2e=residual variance, σ2p=phenotypic variance, h2a=direct heritability c2=ratio maternal permanent environmental variance to phenotypic variance, h2m= maternal heritability; ram= correlation between direct maternal additive genetic effects, h2t=total heritability and SE = standard error
The estimates of total heritability for BWT, WWT and 6WT were 0.21, 0.12, and 0.14, respectively. The total heritability estimates reported by the author were 0.14, 0.12 and 0.21 for Horro sheep (Abegaz, 2002) for BWT, WWT and 6WT, respectively shown little increment across lamb age, which is slightly in contrast with the present results. The result indicated that, maternal effects were important for weights until about 6 months of age.
Based on the best fitted model, the estimate of direct heritability for ADG0-3, ADG3-6 and ADG0-6 were 0.12±0.04, 0.11±0.07 and 0.02±0.05 respectively (Table 5). The estimate indicated that, the inclusion of maternal permanent environmental effects in the analyses can improve the models for daily weights gain traits. The fractions of maternal permanent environmental variance highly reflected for all considered average daily weight gain traits.
The estimate indicates variance due to permanent maternal environmental effects (c2) for ADG0-3(0.21±0.03) and ADG0-6(0.26±0.04) have been found significantly higher than later age daily weight gain traits of ADG3-6 (0.09±0.04). It decreases with increasing lamb age. This could be due to the influences of feeding level at later age of the lambs and the maternal behavior of the dam especially for pre weaning growth traits in the lambs. The value of maternal permanent environmental variance in model (2) for this trait is not significantly different from other models’ values. The result is also found in the range reported by Yacob (2008) for Afar and BHS sheep and lower than the report of Radwan and Shalaby (2017) and Matika (2001) for Rahmani and Sabi sheep respectively.
The estimate of direct heritability for ADG3-6 were comparable with the report of Yacob (2008) that was 0.00 and 0.09 for Afar and BHS sheep under on station management condition.
The estimates of total heritability (h2t) values for ADG0-3, ADG3-6 and ADG0-6 were 0.12, 0.11 and 0.023 respectively that is in similar range to the direct heritability estimates. The estimates are in moderate range with the exception of ADG0-6. Total heritability estimates of ADG0-3 and ADG0-6 is comparable with the finding of Abegaz (2002) for Horro sheep, which is 0.13±0.04 and 0.04±0.03 respectively.Genetic correlations
The present study indicated that BWT had weak genetic correlation with the studied body weights and daily weight gain traits (Table 6). The weak association of BWT with other traits could be due to the fact that BWT is affected by both prenatal and postnatal maternal environments compared to WWT and 6WT traits. It implies that, selection of BWT could not bring positive response to selection on the other traits.
The genetic associations between WWT and other traits, in the present study, were moderate to high except BWT and ADG3-6 months which were low. The negative correlation (-0.35±0.14) between ADG0-3 and ADG3-6 indicating, lambs that grew faster in the pre-weaning period, grew more slowly during post-weaning period and vice versa. Mohammadi et al. (2015) reported similar finding with the present study in Lori sheep breed. WWT and 6WT are the most economically important and easily measured traits and moderate positive genetic correlation was observed between WWT and 6WT (0.52±0.09). The positive genetic correlations between the two traits indicate that the genes that are responsible for increasing WWT result in increasing of 6WT traits. It could be used as selection criteria for improvement in body weights traits.
The positive and moderate genetic correlation that post-weaning body weights and body weight gains may be under the influence of same set of genes (Pleiotropy). WWT and 6WT traits were considered as most appropriate selection criteria in CBBP (Jembere et al., 2016). These genetic correlations were similar to those reported by Safari et al. (2007) and Abegaz (2002) estimated, for Australian Merino and Horro sheep respectively.
The trend was fluctuated and showed a declined trend over the years (Figure 2). It was showed decreasing trend (-0.0026 kg/year and insignificant (P>0.05).When compared to other study, Gizaw et al. (2014a) reported higher and positive genetic change 0.005 kg at 4th generation in BWT for Menz sheep. Since, direct genetic gain for the traits showed slightly negative trend, demonstrate that these traits should not be take into consideration in the selection process by breeder cooperatives.
Perusal of results showed that (Figure 3) direct additive genetic trends had irregular fluctuation and there has been significant (P<0.05) genetic improvement with 0.3 kg in a period of 6-years selection (0.048 kg/year).During the period from 2015 to 2017 the direct additive genetic trend decreased but after this, the direct genetic trend increased in values. The decreasing direct additive genetic trend during 2015 to 2017 may be ascribed to sale of superior breeding rams to areas outside the areas covered by CBBP and inbreeding. This could result for minor decline in growth performance of the flock. The genetic gain 0.048 kg/year was higher than the report for Arman sheep 0.007 kg/year (Mostafa et al., 2011) and lower than the report by Mokhtari, (2010) for Kermani sheep (0.125 kg/year. Gizaw et al. (2014b) reported 0.45 kg of genetic gain at 4th generation of Menz sheep which is higher than the current result of direct genetic gain (0.3 kg). The difference might be due to difference in year of selection, since only 6-year selection data were considered in the present study.
Direct genetic trend over the selection period (2013-18) is presented in Figure 4. The estimated annual direct genetic trend (0.036 kg/year) was positive and highly significant (P<0.01). The fit of the regression shows 73.4% coefficient of determination with the regressed value. The estimate of direct genetic change (0.151 kg) for 6WT provides a good picture of the selection program with respect to 6WT(Table 7),since selection practiced was based on 6WT trait. The present estimate of direct genetic trend was in concurrent with the study of Mokhtari and Rashidi (2010) for Kermani sheep and Mohammadiet al. (2011) for Zandi sheep (0.021 kg/year). Higher estimate was reported by Shaat (2004) in Rahmani sheep (0.135 kg/year).The present estimate is lower than the report of Gizaw et al. (2014b) in Menz sheep 1.3 kg of genetic gain at 4th generation.
Promising results of selection were notified from the ongoing Doyogena sheep CBBP. Direct heritability was medium to high and shows that selection would results in well selection response for growth traits. Total heritability estimates for growth traits were moderate to higher range. The estimate of genetic gain for 6WT trait was the greatest among the body weight traits. It can be concluded that: 6WT could be considered as selection criteria and which could be effective for enhancing growth traits. In order to avoid bias in estimates of genetic parameters, inclusion of maternal effect is important. The genetic progress indicated that, there was satisfactory genetic improvement in most of studied traits due to selective breeding under CBBP in Doyogena breed. Thus, improvement of this breed under CBBP needs to be continued and/or strengthened. Due to the importance of maternal effect, maternal line selection needed to be initiated. The existence of correlations between WWT and 6WT allows an advantage of selection in earlier age. It could also permit culling un-productive lambs in the earlier age.
The authors are thankful to the Southern Agricultural Research Institute (SARI), Areka Agricultural Research Center (AARC) and International Center for Agricultural Research in the Dry Areas (ICARDA) for allowing data and providing technical support for the researcher. We are also grateful to all persons who were involved in this activity for their contribution in different aspects.