The data were from 149 pigs from seven pig genetic groups raised in «Tavriys'ki Svyni» Ltd (Kherson region, Ukraine). The following genetic groups were included in our analyses: LW × LW (n = 19), LW × LN (n = 43), LW × PT (n = 13), LN × LN (n = 15), UM × LN (n = 23), UM × PT (n = 17) and UM×UM (n = 16). The objective this work was evaluation of animal reproductive traits using multivariate analysis. Variables measured and derived included total no. piglets born (TNB), no. piglets born alive (NBA), freq. of stillborn piglets (FSB), total litter birth weight (TLBW), average piglet birth weight (APBW), pre-weaning mortality in piglets (PWM), no. weaned piglets (NW), total weaning weight of litter (TWWL) and average piglet weaning weight (APWW). After standardization, multivariate analyses (Cluster analysis and Principal Component Analysis) were carried out using STATISTICA (StatSoft Ltd.) to place pig interbreeding combinations in groups in accordance with their degree of similarity and verify discriminatory capacity of the original traits in the formation of these groups. The tree diagram showed clear distances between the pig genetic groups studied. In the tree diagram obtained from the analysis of the distances between interbreeding combinations, two distinct groups (clusters) were seen, one with UM × LN and UM × UM animals, and the other with the rest of the pig genetic groups in the study. The eigenvalues for the first two Principal Components (PC1 and PC2) together accounted for near 65% of the variance of the pig’s reproductive traits. The first principal component (PC1) explained 34.9% total variation. It was represented by significant positive loadings for TNB, NBA and TLBW. The second principal component (PC2) accounted for an additional 29.7% of the generalized variance and was represented by significant loadings for NW, TWWL and APWW. Thus, PC1 defined no. piglets and total litter birth weight, while PC2 represented no. weaned piglets and total weaning weight of litter. In conclusion, the multivariate methods (Cluster Analysis and PCA) has been proven to be a very effective method to obtain a synthetic judgment of reproductive traits in pig.
Read full abstract