Evaluation of Resistance Pattern to Fenpyroximateand Pyridaben in Tetranychus Urticae collectedfrom Ggreenhouses and Apple Orchards using Lethal Concentration-slope Relationship

This study aimed to monitor the present and future developments of the resistance of Tetranychus urticae Koch to fenpyroximate and pyridaben, using the relationship of the LC₅₀ and slope of the concentration-mortality line in a probit model, for the provision of reliable resistance management tactics. Tetranychus urticae populations were collected from 16 commercial greenhouses, where various crops were cultivated, as well as from 10 apple orchards throughout Korea. The resistance to fenpyroximate and pyridaben of each population was estimated by calculating the median lethal concentration (LC₅₀), resistance ratio (RR) and slope of the concentration-mortality regression. Most of the greenhouse populations exhibited moderate levels of resistance, whereas the apple orchard populations showed only low levels, indicating that T. urticae populations in greenhouses were more strongly selected than those in apple orchards. Four population groups were established based on either the habitats (greenhouse and apple orchard) or acaricides (fenpyroximate and pyridaben). To test the hypothesis, “the slope is greatest at low and high levels of resistance,” the slopes were regressed as a function of the LC₅₀, and fitted to a polynomial regression. The polynomial regression model explained this relationship well for the four population groups (p < 0.05), indicating that the development of resistance toward fenpyroximate or pyridaben was consistent with the gradient. A laboratory selection study agreed with the results from both acaricide field populations. These results suggest that the gradient was a good indicator of the susceptibility of T. urticae to genetic variations, which was related to the LC₅₀. The application of these findings is also discussed in relation to the resistance management of T. urticae.     

The Use of Multiple Outcomes Analysis in Neurotoxicity: An Example from the New Zealand Study of Offspring of Mothers Exposed to Methyl Mercury

Neurotoxicity has become an important area in the study of the risk to children of noncarcinogens. Many of the tests used to evaluate neurotoxic effects in neurotoxic studies are psychometric tests to assess cognitive, motor, and perceptual performance in individuals in order to determine the presence of psychological problems, impaired educational achievement, or neurological abnormalities. The underlying assumption of these procedures are that a test performance deficit is indicative of a neurological disability, and that such disability derives from the prenatal exposure to an offending substance. The usual analytic paradigm for discerning environmental impacts is to analyze the scores of a particular subtest in a battery taken separately as the dependent variable with regression analysis, which adjusts for cultural background, educational level, and socioeconomic status. The impact of this analysis is to determine the correlation of each of the psychometric test results to exposure. We show a statistical method to simultaneously analyze the multiple outcomes of many tests using General Estimating Equations (GEE) to determine if a correlation exists between the global measure of all these tests and exposure adjusted for covariates of interest and the correlation among the dependent psychometric variables. We show the application of this method to the results of children whose mothers were exposed to mercury during pregnancy.     

Maximum Likelihood Neural Network Prediction Models

Neural networks have received much attention in recent years mostly by non-statisticians. The purpose of this paper is to incorporate neural networks in a non-linear regression model and obtain maximum likelihood estimates of the network parameters using a standard Newton-Raphson algorithm. We use maximum likelihood estimators instead of the usual back-propagation technique and compare the neural network predictions with predictions of quadratic regression models and with non-parametric nearest neighbor predictions. These comparisons are made using data generated from a variety of functions. Because of the number of parameters involved, neural network models can easily over-fit the data, hence validation of results is crucial.     

Prediction of mutations engineered by randomness in H5N1 neuraminidases from influenza A virus

Summary. In this proof-of-concept study, we attempt to determine whether the cause-mutation relationship defined by randomness is protein dependent by predicting mutations in H5N1 neuraminidases from influenza A virus, because we have recently conducted several concept-initiated studies on the prediction of mutations in hemagglutinins from influenza A virus. In our concept-initiated studies, we defined the randomness as a cause for mutation, upon which we built a cause-mutation relationship, which is then switched into the classification problem because the occurrence and non-occurrence of mutations can be classified as unity and zero. Thereafter, we used the logistic regression and neural network to solve this classification problem to predict the mutation positions in hemagglutinins, and then used the amino acid mutating probability to predict the would-be-mutated amino acids. As the previous results were promising, we extend this approach to other proteins, such as H5N1 neuraminidase in this study, and further address various issues raised during the development of this approach. The result of this study confirms that we can use this cause-mutation relationship to predict the mutations in H5N1 neuraminidases. Authors’ address: Guang Wu, Computational Mutation Project, DreamSciTech Consulting 301, Building 12, Nanyou A-zone, Jiannan Road, Shenzhen, Guangdong Province CN-518054, China