Estimates of Breed Differences and Genetic Parameters of Pig Temperament Scores during a Performance Test and its Relationship with Performance Traits

Abstract

YODER, CHAD LEWIS. Estimates of Breed Differences and Genetic Parameters of Pig Temperament Scores during a Performance Test and its Relationship with Performance Traits. (Under the direction of Dr. M. Todd See). The objective of this study was to estimate breed differences for temperament scores during a performance test, and estimate the genetic parameters associated with temperament and performance traits. Chester White (CW), Duroc (D), Landrace (L), and Yorkshire (Y) boars and gilts (n=4,774) were used to estimate breed differences in temperament and genetic parameters. Adjusted backfat (BF), adjusted loin depth (LD), days to 113.4 kilograms (DAYS), estimated percent fat-free lean (LEAN), and three temperament scores: load score (LS), scale score (SS), and vocal score (VS) were recorded. All scores ranged from 1 (calm) to 5 (highly excited). While loading into the scale LS was recorded; SS and VS were recorded while in the scale as ultrasound measurements were taken. Temperament scores were evaluated as categorical traits with statistical models including fixed effects of breed, sex, contemporary group (barn-farm-test date), and off-test weight as a covariate. Order in which pigs were loaded into the scale was included in the model for LS. After initial analysis it was determined VS was best described as two categories, vocal and non-vocal, and was analyzed as a binary trait. Statistical models for BF, LD, DAY, and LEAN included fixed effects of breed, sex, and LS, SS, or VS. Growth rate was adjusted to 113.4 kilograms, while BF and LD were adjusted to the mean weight of the respective breed. Breed differences for LS, SS, and VS were estimated as odds ratios. The odds of increased LS were greater for L (1.62, 1.30; P < 0.01) than D and Y respectively. Landrace had a greater (P < 0.01) probability of a higher SS and VS compared to CW (1.77, 2.37), D (3.31, 3.94) and Y (2.51, 2.46). Yorkshire had greater odds of increased LS (1.25), SS (1.32), and VS (1.60) than D. Chester White had a greater probability of increased LS and SS than D (1.58, 1.87) and Y (1.26, 1.42). Chester White were 1.66 (P < 0.01) times more likely to have a higher VS than D. Genetic models for LS, SS, VS and performance traits were the same as those used to estimate breed differences. However, breed as a fixed effect was excluded and separate analysis was conducted for each breed, and temperament scores were not included in models for performance traits. Variance components were estimated using THRGIBBS1F90 with 150,000 cycles, a 20,000 cycle burn in period, and every 20th sample was stored. Estimates of parameters of interest were obtained as statistics of their respective posterior distributions. Direct heritability estimates were (CW, D, L, Y): LS (0.05, 0.06, 0.06, 0.05), SS (0.22, 0.09, 0.09, 0.10), and VS (0.27, 0.24, 0.37, 0.21). Estimated heritability of LS was low regardless of breed. Heritability estimates for SS, were low with the exception of CW. Heritability estimates for VS were moderate. Direct genetic correlations between SS and VS obtained in a bivariate model were positive and moderate (0.30 to 0.75) indicating that pigs that are more active in the scale are also more vocal. Genetic correlations between temperament traits varied between breed, but were generally moderate and positive. The genetic correlations between LS, SS, and VS with BF, LD, DAYS, and LEAN ranged from - 0.78 to 0.56, and varied among breeds. Based on these results selection for temperament scores collected during a performance test provide potential to alter pig behavior. However, progress in performance traits may be affected if selection for temperament is practiced, while current selection criteria may result in changes in temperament.

Description

Keywords

pigs, temperament, performance, breeds

Citation

Degree

MS

Discipline

Animal Science

Collections