TESTS OF GENETIC STOCK IDENTIFICATION USING CODED WIRE TAGGED FISH

Genetic Stock identification (GSI) uses allozyme variation to determine the composition of mixed-stock fisheries. The GSI method was tested using real fishery data. We report, for the first time in the primary literature, results of tests of GSI in which source stock and ocean-caught mixture samples were separately obtained and the mixture composition was known exactly because the fish used were marked by Coded Wire Tags (CWTs). The accuracy of GSI and its dependence on the quality of genetic data were studied by computer experiments. Rare alleles, which could result from poor sampling procedures, can lead to significant estimation errors. Estimation accuracy depended on the concordance between stocks present in the baseline data and the mixture sample and on the number of loci used in the analysis. Two methods for computing the contributions of groups of source stocks were found to be comparable under most, but not all, conditions. In a blind test of GSI, stock group composition estimates had absolute errors of less than 3%. This suggests that the GSI method can produce accurate stock contribution estimates using real fishery data.