The amino acid compositions of bovine, porcine and fish gelatin were determined by amino acid analysis using 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate as derivatization reagent. Sixteen amino acids were identified with similar spectral chromatograms. Data pre-treatment via centering and transformation of data by normalization were performed to provide data that are more suitable for analysis and easier to be interpreted. Principal component analysis (PCA) transformed the original data matrix into a number of principal components (PCs). Three principal components (PCs) described 96.5% of the total variance, and 2 PCs (91%) explained the highest variances. The PCA model demonstrated the relationships among amino acids in the correlation loadings plot to the group of gelatins in the scores plot. Fish gelatin was correlated to threonine, serine and methionine on the positive side of PC1; bovine gelatin was correlated to the non-polar side chains amino acids that were proline, hydroxyproline, leucine, isoleucine and valine on the negative side of PC1 and porcine gelatin was correlated to the polar side chains amino acids that were aspartate, glutamic acid, lysine and tyrosine on the negative side of PC2. Verification on the database using 12 samples from commercial products gelatin-based had confirmed the grouping patterns and the variables correlations. Therefore, this quantitative method is very useful as a screening method to determine gelatin from various sources.
In-house method validation was conducted to determine amino acid composition in gelatin by a pre-column derivatization procedure with the 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate reagent. The analytical parameters revealed that the validated method was capable of selectively performing a good chromatographic separation for 18 amino acids in less than 40 min; the overall detection and quantitation limit for amino acids fell into ranges of 5.68-12.48 and 36.0-39.0 pmol/μl, respectively; the matrix effect was not observed, and the linearity range was 37.5-1000 pmol/μl. The accuracy (precision and recovery) analyses of the method were conducted under repeatable conditions on different days in random order. Method precision revealed by HorRat values was significantly less than 2, except for histidine with a precision of 2.19, and the method recoveries had a range of 80-115% except for alanine which was recovered at 79.4%. The findings were reproducible and accurately defined, and the method was found to be suited to routine analysis of amino acid composition in gelatin-based ingredients.
A detailed procedure for estimating uncertainty according to the Laboratory of Government Chemists/Valid Analytical Measurement (LGC/VAM) protocol for determination of 18 amino acids in gelatin is proposed. The expanded uncertainty was estimated using mainly the method validation data (precision and trueness). Other sources of uncertainties were contributed by components in standard preparation measurements. The method scope covered a single matrix (gelatin) under a wide range of analyte concentrations. The uncertainty of method precision, μ(P) was 0.0237-0.1128pmolμl(-1) in which hydroxyproline and histidine represented the lowest and highest values of uncertainties, respectively. Proline and phenylalanine represented the lowest and highest uncertainties value for method recovery, μ(R) that was estimated within 0.0064-0.0995pmolμl(-1). The uncertainties from other sources, μ(Std) were 0.0325, 0.0428 and 0.0413pmolμl(-1) that were contributed by hydroxyproline, other amino acids and cystine, respectively. Hydroxyproline and phenylalanine represented the lowest and highest values of expanded uncertainty, U(y) that were determined at 0.0949 and 0.2473pmolμl(-1), respectively. The data were accurately defined and fulfill the technical requirements of ISO 17025:2005.