AN ANALYSIS OF THE DEPENDENCY STRUCTURE BETWEEN A GILT'S PREBREEDING AND REPRODUCTIVE TRAITS. II. PRINCIPAL COMPONENT ANALYSIS

Authors

L. D. Young, Oklahoma Agricultural Experiment Station, Stillwater
R. K. Johnson, University of Nebraska-LincolnFollow
I. T. Omtvedt, Oklahoma Agricultural Experiment Station, Stillwater 74074 and U.S. Department of Agriculture, El Reno, OKFollow

Document Type

Article

Date of this Version

April 1977

Comments

Published in JOURNAL OF ANIMAL SCIENCE, Vol. 44, No. 4 (1977). Copyright American Society of Animal Science. Used by permission.

Abstract

Seventeen variables measured before breeding and three measures of reproduction were taken on 339 purebred Duroc, Hampshire and Yorkshire gilts and 192 two-breed cross gilts resulting from matings among these breeds. Eight principal components accounted for 90% of the dependency structure existing among the 17 traits measured before breeding. Two principal components accounted for 97% of the dependency structure existing among the three reproductive traits.

The first principal component (PCll) from the prebreeding traits was a general measure of growth ability and accounted for 28% of the variation in the 17 measurements. The second principal component (PC12) contrasted slow growing gilts from fast growing litters with fast growing gilts from slow growing litters and accounted for 14.5% of the total variation. The heritability for PC11 was .71 and indicates that selection for gilts with high values for PCll (above .average growth) would be very successful.

The first principal component (PC21) from the reproductive traits contrasted gilts having large numbers of embryos and good embryo survival with gilts having few embryos and poor embryo survival. The second principal component (PC22) contrasted gilts having high ovulation rates and poor embryo survival with gilts having low ovulation rates and good embryo survival. PC21 and PC22 accounted for 57.2% and 39.5%, respectively, of the dependency structure existing between ovulation rate, embryo numbers and embryo survival rate.

Based on the correlations of principal components from growth traits with principal components from reproductive traits, the following conclusions were made: If litter averages are indications of the genetic potential of a gilt selected from that litter, then gilts with a high genetic potential (good litter averages)that exhibit that potential (good individual performance) have high PCll values and are genetically superior for ovulation rate but are genetically inferior for embryo survival rates (high PC22 values). Gilts with a good genetic potential (good litter average) that fail to meet that potential (poor individual performance) have high PC12 values and are genetically inferior for ovulation rate but genetically superior for embryo survival rate (low PC22 values).