Data on precision and accuracy

Regular collaborative trials on amino acid analysis have been conducted for many years by different organizations, and quality assurance of this kind is indeed obligatory today if a laboratory wishes to become accredited, however these results are not published. Organizations that are accessible to all laboratories are AAFCO (USA) with monthly ring tests, Bipea (France) and KDLL (Netherlands) with four to eight samples per year. The results of these and other organizations, which are unfortunately not accessible, provide a more accurate Indication of the state of the art in amino acid analysis than published collaborative trials, which are mostly concerned with method design and standardization. Table 2.2 gives an overview of some published results from literature sources described above, mostly relating to the adoption of official methods by AOAC International or the EU Commission. One of the reasons for the somewhat lower precision of the assay of total concentrations of sulphur-containing amino acids is the prior oxidation in the protein, which introduces additional variation in sample preparation; another reason is due to their almost universally low concentrations of only 0.6-3% in the protein, causing small peaks and greater peak integration errors. Moreover, in IEC cys-teic acid elutes almost in the dead volume of the column, confining matrix interferences to this area. Due to the simplicity of the extraction, analysis of supplemented amino acids gives lower CV than determination of total concentrations. Within-laboratory repeatability CVr are always better by a factor of about 2-5 than the reproducibility CVR between laboratories because the random error is inevitably greater in the latter due to different equipment, handling and calibration. The accuracy and precision of standard methods of amino acid analysis are adequate for the requirements of the feed industry provided that the laboratories are continuously monitored by ring tests and comparisons.

Was this article helpful?

0 0

Post a comment