Date of Award


Degree Type


Degree Name

Master of Science (MS)


Animal Science

Committee Chair

Allan P. Schinckel

Committee Member 1

Michael M. Schutz

Committee Member 2

William M. Muir


The overall objective of this research was to improve pork production economics in United States and China through modeling and genetics. The first study was to estimate the impact of accuracy in which pigs are sorted for market on the optimal market carcass weight (CW) and economic returns. Two types of errors were evaluated, BW estimation error (BWEE) and percentage of pigs not visually evaluated (PNVE). Four levels of BWEE with SD’s of 0, 4, 6 and 8% of BW and four levels of PNVE (0, 8, 16, and 24%) were simulated. Initially, pigs were marketed in three marketing cuts (MCUT), 25% at 169, 25% at 179 and remaining 50% at 193 d of age. The timing of marketing was shifted in seven day intervals. Sort loss was calculated using a market system for a United States pork processor. Sort loss ($/pig) values were fitted to a polynomial function of mean CW for each combination of BWEE and PNVE. In our results, the increase in mean sort loss for each unit increase in CW above 93 kg increased as BWEE and PNVE increased (P < 0.001). With accurate sorting, (BWEE = 0%, PNVE = 0%) the optimal mean age for the 3 MCUT strategy was 190.5 d at a mean CW of 97.0 kg and profit of $3.35/pig. With less accurate sorting (BWEE = 8%, PNVE = 24%), the mean age decreased to 184.5 d with mean CW of 93.4, and profit of $2.00/pig. The optimal market ages and CW’s decreased as BWEE and PNVE increased (P < 0.001). In the second study, we compared selection indexes based on the production costs and economic values from the U.S. and China. Indexes including terminal sire (TSI), maternal line (MLI), and sow productivity (SPI) were calculated based on the production costs and market prices for the U.S. and China. Estimated breeding values (EBV) for: days to 113.5 kg, backfat depth, loin muscle area, number born alive, number weaned, litter weight adjusted to 21 days, days from weaning to estrus, and litter birth weight were provided by a Chinese pig breeding company to evaluate alternative TSI and MLI indexes. The means, SDs, and correlations for the EBV’s and indexes were calculated. The results suggested that the Chinese TSI values were more highly correlated (R = 0.97 to 0.99) with the U.S. indexes than the MLI values (R = 0.92 to 0.97). Overall, the Chinese indexes had greater SD’s (TSI, 58 to 87% greater; SPI, 22 to 26% greater; MLI, 43 to 76% greater). The TSI’s were all highly correlated (R > 0.98) with feed conversion. The Chinese MLI values had greater correlations with TSI and lesser correlations with SPI than the U.S. indexes. The Chinese MLI’s placed greater emphasis on the postweaning traits but less emphasis on sow productivity traits than the current U.S. MLI. The Chinese indexes also tend to select faster growing pigs with less carcass leanness. Feed is the single greatest cost in pork production accounting for approximately 65% of the total production costs. The third study aimed at improving feed efficiency in pork industry via a two-stage selection procedure. We evaluated the impact of adding feed intake records of a percentage of the boars tested on economic improvement in pigs. Data were obtained from a Duroc nucleus herd of a Chinese breeding company including average daily gain (ADG), feed conversion ratio (FG), average daily feed intake (ADFI), days to 115 kg (DAYS), backfat depth (BF), and loin muscle area (LMA). The ADG, FG, and ADFI data were obtained from 570 Duroc boars housed in pens with electronic feeders (FIRE®, Feed Intake Recording Equipment; Osborne Industries, KS, USA). Two sets of estimated breeding values (EBV’s) of ADG, FG, ADFI, DAYS, BF and LMA were estimated by multivariate Best Linear Unbiased Prediction (BLUP), with and without data collected from FIRE feeders. The predicted genetic gain per year was calculated for each trait. Terminal sire indexes (TSI), were calculated based on the production costs and market prices for the U.S. and China. Two-stage selection was applied with the heaviest, sound boars (mean age of 130d) being tested in the FIRE feeder and then top 100 of those boars selected. The TSI value gain estimated with ADFI data was 9% greater than without ADFI data. Estimated genetic gain of FG with ADFI data was 17% greater than without ADFI data. The estimated annual genetic increase in ADFI with ADFI data reduced from 0.041kg to 0.022kg. Overall, there is substantial improvement of the genetic gain of FG when additional individual ADFI records were added on a selected percentage of the boars. In conclusion, the sorting accuracy must be estimated and accounted for in the optimization of market BW’s. The difference regarding production costs, economic value and performance level from different countries or regions impacts the selection objective thus affecting the economic weight assigned to each trait and relative rates of genetic improvement for each trait. Substantial feed efficiency improvement can be achieved by adding ADFI data into BLUP analysis through a two-stage selection procedure. Further studies are needed to develop methods that result in more accurate sorting with a relatively lower cost. For the two-stage selection to improve feed efficiency, the BW in the first stage selection must be measured and accounted for in the BLUP to increase accuracy, and truncation points for each stage of selection need to be evaluated to maximize economic gain over costs associated with measurement.