THE SENSITIVTY OF A TWO-TRAIT SELECTION INDEX TO CHANGES IN ECONOMIC WEIGHTS AND GENETIC PARAMETER ESTIMATES

Abstract

Two-trait selection indexes comprising average daily gain (G) and adjusted average backfat depth (F) were constructed utilizing several permutations of economic and genetic parameters. In these permutations the ratio of economic weights employed (value per kg G: value per mm F) ranged from 5:1 to 75:1, the ratios of heritabilities (G:F) ranged from 0.625:1 to 1:1 and the genetic and phenotypic correlations ranged from zero to 0.25. The relative merits of these indexes were evaluated by comparing the expected genetic merit of progeny produced by the top 10% of the boars identified by each index as applied to a population of 590 boars station tested under the Canadian Record of Performance (ROP) system. Responses were estimated for each component trait (G,F), for feed conversion (FC) and the aggregate genetic response (G + F + FC) in standard deviation units. The net economic value of the aggregate genetic response was calculated as the sum of the values of the genetic responses for the individual traits weighted by their economic values. Variation in the economic and genetic parameters influenced the expected genetic response for the component traits but the aggregate response and its economic value varied little over the range of parameters employed. Indeed, an index constructed from estimates of the actual genetic, phenotypic and economic parameters gave net returns that were only $0.04 per hog greater than an index based on equal economic weights, equal heritabilities and zero correlations. This index, of course, is the simple phenotypic index which leads to the conclusion that results of selection based on a phenotypic index will be fully comparable to those of a highly sophisticated genotypic index. Key words: Selection index sensitivity to parameter variation, genotypic selection index, phenotypic selection index, effectiveness of index selection