Abstract: Determine the efficiency of five models [Linear with two parameters (Ll), Linear with three parameters (L2), second order polynomial (Quadratic), Brody and Richards models to the growth curves of Farafra and Saidi lambs by using monthly records of live weight from birth to 540 days of age were the first goal of this study.The second goal was studying the fixed effects affecting the studied parameters and finally estimates heritabilities for growth curve variables and genetic correlations between these variables.
Data used in the present study were collected from the experimental stations of the Animal Production Research Institute (APRI), Ministry of Agriculture and Land Reclamation, during the years from 1992 to 2005 on two different sheep breeds and locations, Farafra (F) lambs raised at Mallawi experimental station and Saidi (S) lambs raised at Seds experimental station. Farafra data started from year 1992 while Saidi data started from year 1994. A total number of 2559 Farafra and 1539 Saidi lamb records, progeny of 66 and 34 sires, 681 and 318 dams, respectively were available for this study. A Multiple Trait Derivative Free Restricted Maximum Likelihood program (MTDFREML; Boldman et al., 1993) was used to estimate the heritability and genetic correlation with Best Linear Unbiased Prediction (BLUP) methodology.
Least squares analysis of variance option, available in SAS software (SAS, 2004) was used to determine the effect of sex (male or female), type of birth (single, twin or triple), age of dam at lambing which classified into four classes 11 - 29, 30 -44, 45-64 and 65 - 150 month for Farafra breed and 14 - 29, 30 - 43, 44 - 60 and 61 - 124 month for Saidi breed (this classification according to the number ewes on each class, trying to make the number on each class very close to others), year of birth (every year including three successive lambing seasons) and season of birth (February, June or October) on the selected growth curve parameters.
The models were evaluated according to coefficient of determination (R2 values), Akaike's information criterion (AIC) and Bayesian-Schwarz information criterion (BIC). In both breeds, all models fitted the data, with high R2 ranged from 98.32 to 99.82 for over all mean, from 98.65 to 99.85 for males and from 98.79 to 2 99.72 for females. The Brody model gave the best R value of 99.85 in Saidi male, while the Quadratic model gave the lowest R value of 98.32 in Farafra over all mean The AIC and BIC values were lower in Quadratic and Richards models for Farafra and Saidi lambs, respectively than other models.
Absolute Growth Rate (AGR) obtained from L1, L2, Quadratic, Brody and Richards models for Farafra was 0.14, 0.13, 0.10, 0.13 and 0.13 respectively, while for Saidi was 0.13, 0.12, 0.10, 0.11 and 0.12 respectively, in the first month and decreased gradually until reach 0.03, 0.04, 0.04, 0.02 and 0.02 for Farafra and 0.03, 0.04, 0.03, 0.02 and 0.02 for Saidi at 540 days of age, respectively.
Maturing rate obtained from Brody and Richards models for Farafra was 0.0032 and 0.0036 respectively, while for Saidi was 0.0028 and 0.0036 respectively, in the first month and decreased gradually until reach 0.0005 and 0.0005 for Farafra and 0.0006 and 0.0005 for Saidi at 540 days of age, respectively.
Degree of maturing obtained from Brody and Richards models for Farafra was 8.06 and 7.95 respectively, while for Saidi was 6.28 and 6.74 respectively, in the first month and increased gradually until reach 85.96 and 84.63 for Farafra and 81.38 and 84.78 for Saidi at 540 days of age, respectively.
Sex showed significant influence (P<0.0l) on all growth curve parameters except on parameters A and K in Brody model for Farafra sheep that was not significant. Type of birth was significantly influenced (P<0.0l) all growth curve parameters except parameter K in Brody model for Farafra sheep that was not significant. Age of dam significantly influenced (P<0.0l) most growth curve parameters. Year of birth was the major source of variation (P<0.0l) where it significantly affected all growth curve parameters. Season of birth significantly influenced (P<0.0l) all growth curve parameters except parameter A in Brody model for Farafra and Saidi sheep and parameters A and M in Richards model for Farafra sheep that was not significant (P>0.05).
Parameters A, B, C (K) and M were moderately to highly heritable in Farafra lambs (h2=0.24 to 0.74). Genetic correlations between parameters were all positive and ranged from 0.35 to 1.00. In Saidi lambs, parameters had low to moderate heritabilites (0.01 to 0.60). Genetic correlations between parameters ranged from -0.87 to 1.00.
The results of this study suggest that Richards model [Yt = A (1 – Be-KT)m + et ] for growth monitoring can be useful in Farafra and Saidi breeds under both station conditions.
Keys words: growth models, heritability, genetic correlation, Farafra and Saidi lambs, fixed effects.
|