Estimates of genetic parameters for reproductive traits at different parities in Australian hyperprolific large white sows

Heritabilities and genetic correlations were estimated for number of piglets born alive (NBA), average piglet weight at birth (BW) and at 21 days (Wt21d), and number of piglets at 21 days (N21d) for parities 1, 2 and 3 using data from an hyperprolific purebred Large White (LW) herd. Genetic parameters were estimated using residual maximum likelihood methodology under an animal model, using 6,801 records of litters from the first three parities recorded between January 1995 and May 2004. Records from different parities were treated as separate traits. The heritability estimates for parities 1 to 3 were 0.18, 0.17 and 0.19 for NBA, 0.35, 0.34 and 0.41 for BW, 0.23, 0.16 and 0.09 for Wt21d and 0.05, 0.07 and 0.05 for N21d. For the genetic correlations between parities 1 and 2, the only ones that were near to significantly different from unity were NBA (P=0.09) and Wt21d (P=0.1). Between parities 1 and 3, only that for BW was significantly different from unity (P<0.05), while for parities 2 and 3, no correlations were significantly different from unity. Genetic correlations were generally negative between NBA and all three other traits: BW (-0.61, -0.64 and -0.59), Wt21d (-0.36, -0.72 and -0.92) and N21d (-0.23, -0.96 and 0.01). Positive genetic correlations were found between BW and Wt21d (0.63, 0.44 and 0.78) across parities as well as between N21d with weight traits: BW (0.35, 0.33 and 0.37) and Wt21d (0.17, 0.59 and 0.58). The results obtained in this study indicate that enhancing the reproductive performance of hyperprolific sows should be done by selecting other traits as well as litter size. Performance in parity one should be considered as a separate trait for NBA, BW and Wt21d traits.