Sampling for detecting fungicide resistance
L. V. MADDEN (1)
(1) Ohio State University, Wooster, OH, U.S.A.
Phytopathology 101:S228

Detecting shifts in fungicide resistance in fungal populations is of critical importance in developing management strategies in many cropping systems. For some situations, especially where a single gene confers resistance, the shift to resistance in a population can occur rapidly. It is therefore essential to detect a very low prevalence of resistance in fields or in a region while the fungicide is still effective. Sampling methods for rare events provide a framework for precisely determining the upper bound for the prevalence of fungicide resistance, even when all observed isolates are sensitive to the fungicide. That is, when all isolates are sensitive, distributional theory for discrete (binary) variables (with measurement error) shows that there is a nonzero upper bound on the confidence interval for prevalence of resistance. This can be turned around to show the probability of falsely accepting a field (or region) as free of resistant isolates when there is a low (but nonzero) prevalence of fungicide resistance. The sample size can also be determined to declare with 95% confidence that prevalence is below a threshold value when all observed isolates are nonresistant. These concepts can be expanded to develop a two-stage acceptance sampling scheme to detect the presence of resistant isolates in sampled fields based on a sample of isolates from each field.

© 2011 by The American Phytopathological Society. All rights reserved.