Estimates of variance components and heritability using different animal models for growth, backfat, litter size, and healthy birth ratio in Large White pigs

This study compared the estimates of variance components using various animal models for Large White pigs. The traits included three production traits, birth weight (BW), age at 100 kg (Age_100), and backfat thickness at 100 kg (BF_100), and two reproduction traits, number of total born (NTB) and the ratio of healthy births (RHB). Five models including or ignoring common litter environmental effects, maternal effects, and the direct-maternal covariance (sigma(am)) were used for this study. The results showed that the model including all terms, or including all terms except sigma(am), yielded the best-fitting result. The direct variance and heritability were overestimated when the model ignored all previously listed effects, especially for production traits. When all terms were modeled, the direct heritability estimates (h(a)(2) +/- standard error) were 0.038 +/- 0.008, 0.216 +/- 0.022, 0.416 +/- 0.023, 0.066 +/- 0.013, and 0.049 +/- 0.007 for BW, Age_100, BF_100, NTB, and RHB, respectively. The common litter effects reached statistical significance for all traits, and maternal heritability reached statistical significance for three production traits. The direct-maternal correlations were negative for all traits but only reached statistical significance for BW. These results indicate that using a more complex model may result in more accurate estimation of variance components in Large White pigs.