2017.03 Amino Acids (Tryptophan) - Amino-02

2017.03 (JULY 2018) Amino-02 TRP_AOAC_2017.03_MLT_Corrected 080218

FOR ERP USE ONLY DO NOT DISTRIBUTE

the 50-2500 mg/100g RTF range. Three of the matrices had labs that were flagged as outliers using the Cochran’s test and the flagged data was removed for these matrices. Reproducibility (RSD R ) For all product matrices, reproducibility averaged 4.2% RSD with a range of 3.0 - 9.9% RSD. The standard method performance requirements for reproducibility of ≤ 4% RSD for samples with L-tryptophan levels of 5- 50 mg/100g RTF and ≤ 3% RSD for samples with L-tryptophan levels of 50-2500 mg/100g RTF were met for eleven of the fourteen product matrices analyzed. The three product matrices that did not meet the reproducibility requirement were 00729RF00 (Adult Nutritional RTF High Fat), 00730RF00 (Adult Nutritional RTF High Protein), and 410457651Z (Soy-based Partially Hydrolyzed Infant Formula Powder). Reproducibility for 00729RF00 was 4.1% RSD and was slightly above the requirement of ≤ 3% RSD. Reproducibility for 00730RF00 was 4.2% and was also slightly above the requirement of ≤ 3% RSD. The slightly higher reproducibility for these liquid matrices can be attributed to the sample losing homogeneity as it has aged. See figures 1 and 2 for pictures of 00730RF00 and note the large chunks that are easily seen throughout this product. Observations were made from participating labs about the clumping in liquid samples and concerns about this sample’s homogeneity were also noted by the participating laboratories. It should also be noted that the MLT study for AOAC first action method 2015.09 (Vitamin K 1 ) reported similar problems with this sample. Reproducibility for 410457651Z was 9.3% RSD and was above the requirement of ≤ 4% RSD. It should be noted that this sample had an RSDr and RSD iR of 4.95% and 6.92% respectively in the SLV conducted for this method in early 2017. Therefore, it was expected to see a higher reproducibility in this product in the MLT study. Statistical Analysis of MLT Data Using the AOAC International Interlaboratory Study Workbook for Blind (unpaired) Replicates, Version 2.1 to analyze the L-tryptophan MLT data, statistical outliers were identified for three for the fourteen product matrices. The data identified as outliers was removed from the data pool. Data for 410057652Z, 00730RF00, and 00859RF00 were flagged as Cochran outliers. HORRAT values ranged from 0.41 – 1.35 for the fourteen product matrices analyzed. Thirteen of the fourteen product matrices had HORRAT values below 1.00. General Comments There were no major deviations to the MLT protocol during the MLT study. One laboratory did not analyze the practice samples. One laboratory had to re-test day 2 of the study after an error in the preparation of the enzyme was discovered. Once the error was corrected the lab reported acceptable results for day 2. A mistake was made in reporting results for Lab 6. They inadvertently switched samples 00729RF00 and 00730RF00 on day 2 of testing and reported the incorrect results for these two samples. The error was caught and the correct data is reported in the final study report. Analytical columns from multiple vendors were used by the participating labs to complete this MLT study. No significant impact on chromatography was shown using these different column brands.