Selection indexes for improved weaning and marketing body weights of the Burundi local rabbits

Estimates of genetic and phenotypic parameters for growth-related traits involving body weights at 4, 6 (weaning), 8 and 10 (marketing) weeks of age and average daily gains from 4to 6and 6to 10-weeks of age (postweaning daily gain) were computed on 193 males ofthe Burundian local rabbits . Estimates were then used for constructing thirteen selection indexes, involved body weights at 4, 6 and 10 weeks of age and daily gains, to improve body weights of rabbits at weaning and marketing. The use of weaning weight in an index combined with body weight at 4 weeks of age (W„) and posuveaning average daily gain (index,,) or postweaning daily gain solely (index ,2) resulted in a relative efficiency of about 100% as using weaning weight combined with body weights at 4 and 10 weeks of age (index,) or with marketing body weight (index2) or W4 (index3). The accuracy of selection (rT]) using these indexes ranged between 0.890 and 0.906. They lead to a progress in weaning weight from 77-79g and in marketing weight from 54-56g. However, from a practical stand point, selection on index 3, including W4and weaning weight, or on index 5, including weaning weight alone, could be recommended for improving both weaning and marketing weights of Burundian local rabbits.


Introduction
Small scale rabbit production involving minimal inputs, could make a substantial contribution to the supply of animal protein in developing countries.In Burundi, the objective of encouraging production of rabbit meat requires improvement of the productivity of its local breed, a one that is being kept mainly by farmers.To the author's knowledge, the unique attempt was made by the author himself (ANOUS, 2000) to improve meatiness of this breed using selection index.He reported that selection for the füll index involving body weights at 4, 6 and 12 weeks of age (i.e.initial, weaning and marketing body weight, respectively) is expected to lead to increase lean weight (+34.8g)through acheving extra marketing body weight (+56.lg) and higher percentage of lean (+0.62% unit) with maximum accuracy of selection equal 0.87.However, the effect of this method of selection on growth-related traits (i.e.body weights and average daily gains) tili marketing is still needed.Therefore, the purposes of the present study are to quantify the genetic and phenotypic Variation and covariation of growth-related traits in the local rabbits of Burundi, and to calculate the direct and correlated responses expected from selection for both weaning and marketing body weights, considered to be of prime interest to meat rabbit breeders in Burundi, using selection index.

Material and methods
Source of data.The data used in this study were collected during the period from February to June 1995 on 193 male Burundian local rabbits, progeny of 18 unrelated bucks and 46 unrelated does.They were raised at the experimental farm ofthe Institut Superieur d'Agriculture (I.S.A.), Gitega (located approximately 1680m above sea level), Burundi.Local rabbits are of small size and adult males and females weight about 1.7kg.Each doe was contributed 3 to 4 liters with unrepeated mating with the same bück.Management and traits considered.Rabbits were identified at 3 weeks of age and weaned at the age of 6 weeks.From weaning up to marketing at 10 weeks of age, all animals were fed ad libitum on a 16 % crude protein commercial pelleted ration containing 2600 Kcal digestible energy/kg diet.Roughages and green grass were also provided.The body weights in grams at 4 (W 4 ), 6 (W 6 ), 8 (W 8 ) and 10 (W )0 ) weeks of age were recorded for each rabbit.Average daily gain from 4-to 6-weeks of age (DG 4 .6 ) and postweaning average daily gain from 6-to 10-weeks of age (DG 6 .io)were calculated (g/d).Statistical analysis.Heritabilities and genetic and phenotypic correlations were estimated from the sire components of variance and covariance by the Least Squares and Maximum Likelihood program (HARVEY, 1990) aecording to the following model: = u + S i + D ij + E ijK , = the weight or average daily gain measured on the k th animal ofthe j th dam nested within the i th sire; = the population mean for the trait Y; = the random effect ofthe i th sire (i = 1,2 , .. .,18);= the random effect of the f* dam nested within the i" 1 sire (j = 1,2, ..., 46); = the random error assumed NID~(0, o 2 E).Estimates of heritability from the maternal half-sib correlations were less than zero for some traits.Therefore, only estimates calculated by paternal half-sib analysis were considered in the present study.As parity and litter size were not significant in a preliminary analysis ( P>0.10), they were ignored in the final one .Both weaning and marketing body weights were retained in the aggregate genotype with equal economic value (assumed to be 1).Body weights at 4 , 6 (weaning) and 10 (marketing) weeks of age together with daily gains from 4-to 6-and from 6-to 10weeks of age were used in different combinations to compute thirteen selection indexes (CUNNINGHAM et al., 1970) to select for increased body weights at weaning and marketing.

Results and discussion
Heritability estimates.The value of heritability estimates (h 2 ) for growth-related traits together with their Standard errors (S.E.) are shown in Table 1.
Table 1 Mean (X), phenotypic Standard deviation (a P ) and heritability (h 2 ) with its Standard error (+S E) for live The value of h estimate for body weight increased with increasing age from 0 67 at 4week of age up to 0.98 at 8-week of age, then it decreased.In agreement with the conclusion of previous reports reviewed by KHALIL et al. (1986), estimates of heritability for body weight were highest at younger ages, declining to the lowest values 2 after weaning and increasing again at the older ages.The h -value for body weight was higher at weaning than marketing (0.81 vs 0 67) This agrees with the findings of MOSTAGEER et al. (1970) on Giza White rabbits (0.185 vs 0.150) and KHALIL (1986) on Bouscat (0.46 vs 0.24) and Giza White (0.65 vs 0.39) breeds.The estimates of sire-component h 2 show that selection for body weight at earlier^ages may be useful method for improving early rabbit growth.Generally, the h values obtained in the present study for the body weight at 4, 6, 8 and 10 weeks of age were higher than those reported in the literature (0.10 to 0 55 at 4 weeks; MOSTAGEER et al., 1970;OUHAYOUN et al., 1973;EL-AMIN 1974' CHEVALET, 1976;VRILLON et al., 1979;0.185 to 0.650 at 6 weeks;MOSTAGEER et al., 1970;ALVAREZ et al., 1974;KHALIL, 1986;FERRAZ et al., 1991-0 20 to 0.72 at 8 weeks; MOSTAGEER et al" 1970;EL-AMIN, 1974;KHALIL, 1986-0 15 to 0.56 at 10 weeks; MOSTAGEER et al., 1970;OUHAYOUN et al 1973-VRILLON et al., 1979;KHALIL, 1986;FERRAZ et al., 1991).This reflects the high contribution ofthe genetic additive variance to the total variance for body weights in this breed of rabbit compared to the other breeds.The heritability estimates for the average daily body weight gain were moderate to high, but lower than those of body weight.They were higher between weaning and marketing ^than between4-week of age and weaning (0.53 vs 0.47).However, the ränge of h values reported in the literature for these traits (0.13 to 0.31; MOURA et al., 1991 andKROGMEIER et al., 1994, respectively) indicates low to moderate contribution of the genetic additive variance to the total variance.Thus, using body weights for constructing selection indexes to increase both weaning and marketing body weights of rabbits in the present study will lead to more gain than using daily gains.Genetic and phenotypic correlations.Estimates of genetic and phenotypic correlations among various growth-related traits are given in Table 2.   Table 2 .i Genetic (above diagonal) and phenotypic (below diagonal) correlations among the live Performance traits (Genetische (oberhalb der Diagonalen) und phänotypische (unterhalb der Diagonalen) Beziehungen zwischen

-
The genetic correlations among body weights were generally higher than the phenotypic correlations.They were all positive, so that rabbits more heavy at given age tended also to be more heavy at any later age.The lowest values obtained were between body weight at 4-week of age and marketing body weight (r G = 0.79 and r P = 0.63).This agrees with the conclusion reported by KHALIL et al. (1986).Thus, rabbit body weight at earlier ages can be used for selection to improve its body weight at later ages.The genetic correlations between body weights and average daily gains were also higher than the phenotypic correlations.Genetically and phenotypically, the relation was positive among body weights and average daily gain from 4-to 6-weeks of age, while negative with average daily gain from weaning to marketing.This Opposition in sign may indicate that, in rabbits, body weight before weaning is subjected to a large maternal influence, while it was influenced after weaning by the environmental factors (ex.ration, climat, etc).Body weight at 6 weeks of age was highly correlated both genetically and phenotypically to body weight at 4 (r G = 0.90 and r P = 0.81) and 10 (r G = 0.94 and r P = 0.82) weeks of age, indicating that weaning weight could replace alone the two other body weights (W 4 and W )0 ) as an index trait.The basic traits in the aggregate genotype (W 6 and W, 0 ) showed large and positive genetic relationships with the other growth-related traits; the relationship was negative with postweaning average daily gain (r G = -0.98 and -0.77, respectively).This indicates that selection for increasing body weights at weaning and marketing should affect positively body weights at other ages and negatively postweaning daily gain.Selection indexes.Thirteen selection indexes arranged in three strategies (i, ii and iii), aecording to the use of one or a combination of body weights and average daily gains, were construeted.The weighing factors (b-values) and correlations ofthe indexes with the total aggregate genotype (r j-values) are given in Table 3.Using body weights alone in an index (strategy i) showed that selection on index 1, including W 4 ,W 6 and Wi 0 is theoretically capable of predicting the aggregate genetic value of an animal 0.90 as accurately as a perfect index (r-rr = 0.899).Assuming the efficiency of index 1 to be 100 •%, the Omission of W 4 or W, 0 from index 1 (indexes 2 and 3, respectively) resulted in practically the same value of m as index 1 (r TI = 0.899 and 0.890, respectively).The relative efficiency of the reduced indexes including W 6 , W| 0 or W 4 alone (indexes 5, 6 and 7) were 97.89, 89.21 and 78~87 %, respectively.Using index 5, including W 6 solely, leads practically to the same value of r T i as with indexes 1,2 and 3 (r T( = 0.880).
In terms of selection efficiency , the use of average daily gains solely in an index (strategy ii) would not be as promising as the use of body weights alone (strategy i) or body weights combined with daily gains (strategy iii).Using W 6 combined with W 4 and postweaning daily gain (index 11) or with postweaning daily gain (index 12) resulted in a relative efficiency of about 100 % as using indexes 1, 2 and 3. From a practical Standpoint, selection on index 3 including W 4 and W 6 or on index 5 including W 6 alone could be considered as the best for improving both weaning and marketing body weights of Burundian local rabbits since their application is earlier compared to the use of indexes 1 and 2 (involving W| 0 measured at later stage ofthe animal life), and easier compared to the use of indexes 11 and 12 ( involving postweaning daily gain).The effect ofthe different indexes was calculated as the expected improvement in both weaning and marketing body weights as well as correlated response in the other growth-related traits when selection on indexes 1 to 13 and the results are given in Table 4. Generally, selection using body weights alone (indexes 1 to 7) or combined with postweaning daily gain (indexes 11 to 13) gave the greatest progress in marketing body weight, especially when W 6 was involved (from 42 to 56g), while selection using daily gains alone (indexes 8 to 10) gave the lowest improvement (from 30 to 40g).On the other hand, selection using daily gains gave the greatest progress in weaning weight (from 61 to 102g), especially when DG 4 ^ was involved , however selection using body weights alone or combined with daily gains gave the same improvement in weaning weight (from 63 to 79g).The decreasing in post weaning average daily gain of rabbits is much less when daily gains alone in an index (from -0.3 to -0.6 g/d) compared to the use of body weights alone (from-0.5 to -0.8 g/d) or body weights combined with daily gains (from -0.7 to -0.8 g/d).Selection on all indexes is expected to increase all the other growthrelated traits differently.The index incorporating body weights at 4, 6 and 10 weeks of age (index 1) gave the greatest progress in marketing body weight (+56 g) as well as the reduced index involving body weights at weaning and marketing (index 2), however a moderate progress in weaning weight (+77 g) was observed with their use.On the other hand, the index incorporating DG 4 -6 alone (index 9) or together with postweaning daily gain (index 11) gave the greatest progress in weaning weight (+95 and +102 g, respectively) with a small progress in marketing body weight (+30 and +40 g, respectively).The indexes 3, 5, 11 and 12 gave about the same improvement in both weaning and marketing weights , and also the same improvement in the other growthrelated traits.The magnitude of the amount of progress in growth-related traits is comparable with this obtained using index 1 or 2. Therefore, using one or more than one of the above mentioned indexes would lead to the same progress in growth-related traits.Taking into consideration the expected genetic changes in growth-related traits as well as in weaning and marketing body weights, it appears more practical to the Burundian farmer to use the index including weaning weight alone (index 5) or combined with body weight at 4-week of age (index 3) for improving efficiently both weaning and marketing body weights of rabbits since their application is earlier.The use of these

Table 3
Weighing factors (b-values) and accuracy of selection in absolute (R TI ) and relative (RE) values in indexes used to improve both weaning and marketing body weights (Wichtungsfaktoren (b-Werte), absolute (R T i) und relative

Table 4
Expected genetic changes in growth-related traits when using indexes to improve both weaning and marketing body weights ( intensity of selection = 1.00 ) (Erwarteter genetischer Gewinn in den Wachstumsmerkmalen bei unterschiedlichen Strategien in den verschiedenen Indices)