Sexual maturity stages of canned chum salmon segregated using static headspace volatile patterns

Abstract Forty-four volatile compounds in canned chum salmon ( Oncorhynchus keta ) were isolated by static headspace gas chromatography and used in multivariate statistical procedures to discriminate three stages of sexual maturity. Principal component analysis was initially carried out; eight principal components (PC) were selected and accounted for 84% of the variation in the volatiles. The first two PC mainly contained extraneous variation while PC3–PC6 correlated with various volatiles substantive to differentiating stages of maturity. Although the application of linear (LDA), quadratic (QDA), and non-parametric (NPDA) discriminant analyses to PC scores all separated the groupings with high correct classification rates (95–96%), NPDA, circumvented the violation of parametric assumptions and, thererfore, constituted the method of choice for classifying canned chum salmon of unknown sexual maturity.