Treatment of Data for Duplicated Gene Loci in Mixed-Stock Fishery Analysis

Abstract

Isoloci (pairs of duplicated loci with a predominance of shared alleles of identical electrophoretic mobility) are common in salmonid fishes and present special problems for the analysis of mixed-stock fisheries. The method currently used to treat isolocus data (which assumes that allele frequencies at both loci are the same) leads to bias if the true frequencies (P, Q) are not identical. Using simulated mixtures drawn from populations with (1) hypothetical allele frequencies and (2) allele frequencies observed at four isolocus systems in samples of chinook salmon from the Fraser River system, we show that a better method is to use the maximum likelihood estimates ([Formula: see text], [Formula: see text]) of allele frequencies at the individual loci. In simulations in which the population frequencies at the two loci of an isolocus pair differed, use of ([Formula: see text], [Formula: see text]) resulted in less error in estimating the true mixture proportions than did use of a single average frequency, whereas the two methods produced very similar results if population frequencies were identical at both loci.