This year the amino acid analysis subcommittee sent three types of samples to member facilities: a control of unspecified protein, the same protein plus Tris, and the same protein adsorbed to PVDF. We asked that each sample be run in triplicate with normal care, expecting thereby to receive representative results with which to assess both average and achievable precision and accuracy. We also asked about instrumentation, methods, and special procedures. Facilities favored vapor- to liquid-phase hydrolysis about 2:1, all use 6N HCl, most use 105115deg.C. Pre-column methods, predominantly using PITC, were slightly favored over post-column methods (ion-exchange), mostly with ninhydrin detection. Average compositional error (inaccuracy) was 13% for the control sample, 20% for the sample containing Tris, and 27 % for the PVDF sample, reflecting the increasing degree of difficulty. While the most successful sites achieved 95% accuracy for both of the first two samples, these same sites experienced about three times higher error for the samples on PVDF. Reproducibility and recovery of protein followed a pattern similar to percent error. In general there were no significant differences in the results from pre- and post-column methods and both provided many excellent analyses. However, certain situations favored one over the other or required different procedures to obtain good results. For instance, the Tris-containing sample had to be desalted for pre-column methods to work well, while post-column methods had more difficulty quantifying proline. Handling losses and sample contamination were common sources of error but not universal. Also, calibration was not always straightforward and successful. The most difficulty was experienced with PVDF samples--these methods require further improvement to obtain results beyond a crude estimation of protein. About a fourth of the sites identified the protein correctly as horse apomyoglobin.
Return to the The ABRF Home Page
Created: 15th August 1995
Last modified: 15th August 1995