Accuracy of amino acid analysis of fish meals by ion-exchange and gas chromatography

Variations in amino acid content of fish meals can be due to differences between meals or to analytical shortcomings. The purpose of the cooperative study reported here was to compare the accuracy of the ion-exchange (IE) and gas chromatographic (GC) methods when applied to fish meal. The between-laboratory and within-laboratory variations were determined by analysing hidden duplicates of eight meals in four laboratories using IE and in a further three laboratories using GC. The work exposed certain pitfalls and inconsistencies in replicating results. A major source of variation in the IE method was identified as the instability of the ninhydrin reagent leading to the consequent variation in the colour yield of some amino acids which was not compensated by the use of norleucine as an internal standard. Significant between-laboratory variations and fish meal X laboratory iteractions for certain amino acids existed. The work confirmed a need for replicate analysis of samples, whether done by IE or GC. Without discarding any outlying values, coefficients of variation for repeatability (within-laboratory variation) varied from 3.5 to 9.7% for IE and 3.5 to 8.4% (histidine at 20%) for GC. Coefficients of variation for reproductability (between-laboratory variation) varied from 4.1 to 14.4% for IE and 4.5 to 11.9% (20% for histidine) for GC. The values for lysinereproductability were 10.8% and 4.8% and for methionine were 13.6% and 8.2% respectively for IE and GC. Neither of the methods was clearly superior for determining the amino acid content of fish meals. It would appear, however, that the GC method tended to result in lower within-laboratory variance, with arginine and histidine as exceptions. An amino acid mixture, a centrally prepared hydrolysate and a protein of known composition were also studied.