Spatial distribution of single-nucleotide polymorphisms related to fungicide resistance and implications for sampling.

Spatial distribution of single-nucleotide polymorphisms (SNPs) related to fungicide resistance was studied for Botrytis cinerea populations in vineyards and for B. squamosa populations in onion fields. Heterogeneity in this distribution was characterized by performing geostatistical analyses based on semivariograms and through the fitting of discrete probability distributions. Two SNPs known to be responsible for boscalid resistance (H272R and H272Y), both located on the B subunit of the succinate dehydrogenase gene, and one SNP known to be responsible for dicarboximide resistance (I365S) were chosen for B. cinerea in grape. For B. squamosa in onion, one SNP responsible for dicarboximide resistance (I365S homologous) was chosen. One onion field was sampled in 2009 and another one was sampled in 2010 for B. squamosa, and two vineyards were sampled in 2011 for B. cinerea, for a total of four sampled sites. Cluster sampling was carried on a 10-by-10 grid, each of the 100 nodes being the center of a 10-by-10-m quadrat. In each quadrat, 10 samples were collected and analyzed by restriction fragment length polymorphism polymerase chain reaction (PCR) or allele specific PCR. Mean SNP incidence varied from 16 to 68%, with an overall mean incidence of 43%. In the geostatistical analyses, omnidirectional variograms showed spatial autocorrelation characterized by ranges of 21 to 1 m. Various levels of anisotropy were detected, however, with variograms computed in four directions (at 0°, 45°, 90°, and 135° from the within-row direction used as reference), indicating that spatial autocorrelation was prevalent or characterized by a longer range in one direction. For all eight data sets, the ß-binomial distribution was found to fit the data better than the binomial distribution. This indicates local aggregation of fungicide resistance among sampling units, as supported by estimates of the parameter θ of the ß-binomial distribution of 0.09 to 0.23 (overall median value = 0.20). On the basis of the observed spatial distribution patterns of SNP incidence, sampling curves were computed for different levels of reliability, emphasizing the importance of sample size for the detection of mutation incidence below the risk threshold for control failure.

Spatial Pattern of Strawberry Powdery Mildew (Podosphaera aphanis) and Airborne Inoculum.

The relationship between strawberry powdery mildew and airborne conidium concentration (ACC) of Podosphaera aphanis was studied using data collected from 2006 to 2009 in 15 fields, and spatial pattern was described using 2 years of airborne inoculum and disease incidence data collected in fields planted with the June-bearing strawberry (Fragaria × ananassa) cultivar Jewel. Disease incidence, expressed as the proportion of diseased leaflets, and ACC were monitored in fields divided into 3 × 8 grids containing 24 100 m2 quadrats. Variance-to-mean ratio, index of dispersion, negative binomial distribution, Poisson distribution, and binomial and beta-binomial distributions were used to characterize the level of spatial heterogeneity. The relationship between percent leaf area diseased and daily ACC was linear, while the relationship between ACC and disease incidence followed an exponential growth curve. The V/M ratios were significantly greater than 1 for 100 and 96% of the sampling dates for ACC sampled at 0.35 m from the ground (ACC0.35m) and for ACC sampled at 1.0 m from the ground (ACC1.0m), respectively. For disease incidence, the index of dispersion D was significantly greater than 1 for 79% of the sampling dates. The negative binomial distribution fitted 86% of the data sets for both ACC1.0m and ACC0.35m. For disease incidence data, the beta-binomial distribution provided a good fit of 75% of the data sets. Taylor's power law indicated that, for ACC at both sampling heights, heterogeneity increased with increasing mean ACC, whereas the binary form of the power law suggested that heterogeneity was not dependent on the mean for disease incidence. When the spatial location of each sampling location was taken into account, Spatial Analysis by Distance Indices showed low aggregation indices for both ACCs and disease incidence, and weak association between ACC and disease incidence. Based on these analyses, it was found that the distribution of strawberry powdery mildew was weakly aggregated. Although a higher level of heterogeneity was observed for airborne inoculum, the heterogeneity was low with no distinct foci, suggesting that epidemics are induced by well-distributed inoculum. This low level of heterogeneity allows mean airborne inoculum concentration to be estimated using only one sampler per field with an overall accuracy of at least 0.841. The results obtained in this study could be used to develop a sampling scheme that will improve strawberry powdery mildew risk estimation.

Quantification of Airborne Inoculum as an Aid in the Management of Leaf Blight of Onion Caused by Botrytis squamosa.

Botrytis leaf blight, caused by Botrytis squamosa, is a common and frequently damaging disease of onion crops, but the severity of epidemics varies widely from year to year. The disease is initiated and spread by airborne conidia. The relationship between airborne conidium concentration (ACC) and lesion development was studied in the field. A linear relationship was found between ACC and number of lesions per leaf: ACC values of 10 to 15 and 25 to 35 conidia m-3 were associated with 1 and 2.5 lesions per leaf, respectively. In 2000 and 2001, at three sites, four different criteria were used to start a fungicide spray program and their effect on epidemic development was compared with that of a grower's conventional schedule. The criteria were: at the fourth-true-leaf growth stage; according to an inoculum production index; when the ACC reached 10 to 15 conidia m-3; and when the ACC reached 25 to 35 conidia m-3. A nonsprayed control plot was included in the trial. Fungicide programs started when the ACC reached 10 to 15 conidia m-3 were as effective as the conventional program, but used fewer sprays. A fungicide spray program based on measurements of ACC and disease severity was evaluated in 2002 and 2003 in five and three commercial onion fields, respectively. At each site, half of the field was sprayed according to the grower's schedule and, in the other half, fungicide sprays were initiated when a threshold of 10 to 15 conidia m-3 or five lesions on the lower leaf (whichever came first) was reached. Overall, the number of fungicide applications was reduced by 75 and 56% in 2002 and 2003, respectively, without causing significant yield reduction. In both years, the reduction in number of fungicide applications was due mainly to the delay in initiation of the fungicide program.