Use of experimental designs with quantitative ELISA.

Precise use of enzyme-linked immunosorbent assay (ELISA) as a quantitative technique depends on repeatability of color development and its measurement. Variation in measured response among wells, within and among microtiter plates, often precludes such precision. For example, plates with all wells treated uniformly exhibited unacceptable optical density differences in excess of 0.35 and 0.25 O.D. U among row and column averages, respectively. Arrangement of samples on plates according to classical experimental designs, with compact blocking features and two-dimensional control over spatial patterns, provides a possible remedy. Analysis of variations over uniformly treated plates demonstrated the potential for increased precision when such designs are used instead of random arrangements. Retrospective analysis of more than 100 tests performed with various experimental designs confirmed that this potential was realized when using Youden square and lattice square designs. Several designs appropriate for microtiter plates are presented and their conduct described.