(X - l)2 + 2nx where X is the dilution factor in the series and n is the subjective precision of picking the matching mimic in the series. Ideally, n could be assumed to be 2, but often, especially when the dilution factor is less then 2, one has to take n = 1. Analogy with a ruler is helpful for explanation: If the smallest grid on a metric ruler is 1 mm, one can assume that the best measurement with such ruler is within 1/2 mm (rounding to the nearest mark). This corresponds to n = 2 and is probably justified if carrying out three-, four-, and tenfold dilutions of mimics, when relative changes of concentrations in the series are large, and differences are easy to see; n = 1 in the ruler analogy is measuring to within 1 mm (marks are hard to see, or the boundary of a measured point is not sharp). For X = 10 or 2 (as suggested in the CLONTECH protocol), the error calculated this way is 99% or 60% (n = 1) and 82% or 33% (n = 2), respectively.
The "symmetrically" corrected value itself is calculated as C = C (1+(X - Ifflnx) (correction term underlined). Obviously, the precision of measurements with the mimic protocol is inherently limited: It can be increased by making smaller dilutions, but then it is harder to pinpoint the matching dilution among similar neighbors. The best that can be achieved in practice is probably 30%.
Was this article helpful?