Comparison of interlaboratory variation in amino acid analysis and rat growth assays for evaluating protein quality

Estimates of inter- and intralaboratory variation of protein efficiency ratio (PER), relative PER (RPER), net protein ratio (NPR), relative NPR (RNPR), and nitrogen utilization (NU) were compared with those of amino acid analysis in the same batches of 7 protein sources (ANRC casein, egg white solids, minced beef, soy assay protein, rapeseed protein concentrate, pea flour, and whole wheat flour). Interlaboratory variation (estimated as between-laboratories coefficients of variation, CV) of NPR and RNPR (up to 6.0%) was lower than that of PER (up to 20.2%) and RPER (up to 18.5%). The interlaboratory determination of NPR and RNPR was also more reproducible than that of most essential amino acids (CV up to 10.0%), especially tryptophan (CV up to 23.7%), cystine (CV up to 17.6%), and methionine (CV up to 16.1%). Intralaboratory variation (estimated as within-laboratories CV) of amino acid analysis (up to 4.7%), however, was comparable to that of protein quality indices in most protein sources (up to 6.0%). The significant (P less than 0.01) positive correlations (r = 0.68-0.74) between amino acid scores and protein quality indices based on rat growth were further improved when amino acid scores were corrected for digestibility of protein (r = 0.73-0.78) or individual amino acids (r = 0.79-0.82).