My research focuses on the development of a novel method for the fast detection of aflatoxin-producing fungi from the volatile organic compounds that they produce. Aflatoxins have received great attention because of their demonstrated potent carcinogenic effect in susceptible laboratory animals and their acute toxicological effects in humans. Traditional detection and quantification techniques are considered time-consuming, high cost, and require technical professionals. The `odor' or so called volatile metabolites released by a fungus is the key for fast detection. Several researchers have reported that diverse fungi species have unique volatile metabolite patterns. This study focuses on answering several questions: Is it possible to discriminate aflatoxins-producing fungi from other fungi based on volatile metabolites? What are the key discriminating biomarkers related to each fungus? Does the growth environment have an effect on the production of volatile metabolites? What chemicals are consistently emitted by a fungus under varied conditions? To answer these questions, one toxigenic and one nontoxigenic A. flavus isolate were studied to evaluate the microbial volatile organic compound (MVOC) profiles. The results described in chapter two of this dissertation indicate that MVOC production is time-dependent and that aflatoxigenic and non-aflatoxigenic strains have different MVOC expression patterns. Chapter three describes the effects of experimental parameters on fungal volatile metabolites. The identity and quantity of MVOCs can be affected by many factors including SPME fiber type, fungal growth media, and growth temperature. A CAR/PDMS coated fiber performed better than the other SPME fibers by collecting a larger variety and quantity of MVOCs. Fungi grown on the chemical defined liquid media produced much larger quantities of MVOCs compared to the other media. The highest MVOC production results were found at 30 degrees Celsius. The fungi discrimination study was extended in chapter four by including 3 toxigenic and 3 non-toxigenic isolates using multivariate analysis. The results indicate that volatile patterns vary even at the fungal isolate level and that discrimination of aflatoxin-producing fungi from non-toxigenic fungi is possible.

No description available.

The cosmopolitan Pteridium aquilinum (L.) Kuhn is widespread throughout eastern North American, where it is represented primarily by Tryon's (1941) var. latiusculum (Desv.) Underw. and var. pseudocaudatum (Clute) Heller. The taxonomy of Pteridium is controversial. Fourteen isozyme loci and 12 morphological characters were used to assess the taxonomic relationship of these two varieties. Isozyme data indicated a high mean genetic identity (I = 0.976) between eleven bracken populations. Strong patterns of geographic variation for isozyme allele frequencies were also observed. The isozyme results did not separate the two taxa. Numerical analysis of the morphology distinguished the two taxa when the qualitative characters were used alone or in conjunction with some of the quantitative traits. All qualitative characters differed significantly between the two taxa. No perceptible geographic pattern of variation was observed. Morphological distinctiveness was maintained even in those localities where both taxa were present, with few or no intermediates being found. Isozyme evidence suggestive of gene flow between the two varieties was found at Greensboro, NC, where the two morphotypes were easily recognizable. The isozyme evidence strongly indicates conspecificity, while the morphological evidence supports their status at the varietal level. / Master of Science

No description available.

No description available.

As a useful tool for multivariate analysis, sufficient dimension reduction (SDR) aims to reduce the predictor dimensionality while simultaneously keeping the full regression information, or some specific aspects of the regression information, between the response and the predictor. When the goal is to retain the information about the regression mean, the target of the inference is known as the central mean space (CMS). Ordinary least squares (OLS) is a popular estimator of the CMS, but it has the limitation that it can recover at most one direction in the CMS. In this dissertation, we introduce two new estimators of the CMS: the sliced OLS and the hybrid OLS. Both estimators can estimate multiple directions in the CMS. The dissertation is organized as follows. Chapter 1 provides a literature review about basic concepts and some traditional methods in SDR. Motivated from the popular SDR method called sliced inverse regression, sliced OLS is proposed as the first extension of OLS in Chapter 2. The asymptotic properties of sliced OLS, order determination, as well as testing predictor contribution through sliced OLS are studied in Chapter 2 as well. It is well-known that slicing methods such as sliced inverse regression may lead to different results with different number of slices. Chapter 3 proposes hybrid OLS as the second extension. Hybrid OLS shares the benefit of sliced OLS and recovers multiple directions in the CMS. At the same time, hybrid OLS improves over sliced OLS by avoiding slicing. Extensive numerical results are provided to demonstrate the desirable performances of the proposed estimators. We conclude the dissertation with some discussions about the future work in Chapter 4. / Statistics

Thesis. 1977. M.S.--Massachusetts Institute of Technology. Dept. of Earth and Planetary Sciences. / Microfiche copy available in Archives and Science. / Bibliography : leaves 52-54. / M.S.

Multivariate statistical methods have been strongly recommended in behavioral research employing multiple dependent variables. While the techniques are readily available, there is still controversy as to the proper use of the methods that have been developed for analyzing and interpreting data after finding a significant pairwise difference with a multivariate analog of the two group t-test, known as Hotelling's T². A Monte Carlo simulation was conducted to investigate the relative power of four post-MANOVA tests under violations of multivariate homoscedasticity and normality. The four methods for analyzing multivariate group differences following a significant Hotelling's T² were: (1) univariate F; (2) Bonferroni; (3) multiple Bonferroni; and (4) simultaneous F. Depending on the conditions examined, either the univariate F test or the multiple Bonferroni procedure was shown to be the most powerful for detecting a true difference between two groups. The following are the major conclusions drawn from the investigation: (1) Power levels of post-MANOVA tests remain constant under violations of multivariate normality, however, they change considerably in the presence of heterogeneity; (2) The univariate F test provides the most liberal power levels and the simultaneous F test provides the most conservative, regardless of sample size, effect size, distribution shape, and degree of violation; (3) As the size of the effect increases, the rate of correct rejections of a false null hypothesis increases; (4) As sample size increases, the rate of correct rejections of a false null hypothesis increases; (5) Regardless of heterogeneity level, power is always larger at larger group size levels; and (6) Within each group size level, power decreases as heterogeneity increases. Analytical comparisons show simultaneous F tests have the least power, Bonferroni methods to be intermediate, and univariate F tests most powerful under violations of multivariate heterogeneity. / Ph. D.

Linear discriminant analysis is the classification of an individual as having arisen from one or the other of two populations on the basis of a scalar linear function of measurements of the individual. This paper is a population and large sample study of linear discriminant analysis. The population study is carried out on three levels: (1.1) (a) with loss functions and prior probabilities, (b) without loss functions but with prior probabilities, (c) with neither. The first level leads to consideration of risks which may be split into two components, one for each type of misclassification, i.e. classification of an individual into population I given it arose from population II, and classification of it into II given it arose from I. Similarly, the second level leads to consideration of expected errors and the third level leads to consideration of conditional probabilities of misclassification, both again which may be divided into the same two components. At each level the "optimum" discriminator should jointly minimize the two probability components. These quantities are all positive for all hyperplanes. Either one or any pair may be made equal to zero by classifying all individuals of a sample into the appropriate population; but this maximizes the other one. Consequently, joint minmization must be obtained by some compromise, e.g. by selecting a single criterion to be minimized. Two types of criteria for judging discriminators are considered at each level: (1.4) (i) Total risk (a) (1.5) Total expected errors (b) (1.6) . Sum of conditional probabilities of misclassification (c) (1.7) (ii) Larger risk (a) (1.8) Larger expected error (b) (1.9) Larger conditional probability of misclassification (c). These criteria are not particularly new, but have not been applied to linear discrimination and not been all used jointly. If A is a k-dimensional row vector of direction numbers, X a k-dimensional row vector of variables, and a constant, a linear discriminator is (1.10) AX' = o, which also represents a hyperplane in k-space. An individual is classified as being from one or the other population on the basis of its position relative to the hyperplane. The parameters A and c ot (1.10) were investigated to find those sets of values which minimize each of the two criteria at various levels. Exact results were found for A under some circumstances and approximate results in others. At the levels (b) and (c), when exact results were obtained, they were the same for both criteria and were independent or c. Investigation of the c’s showed the c’s to be exact functions of A and the parameters and yielded one c for each criterion. At level (c), the c's for criteria (i) and (ii), c(min) and c(σ), respectively, were compared to c(m), a population analog of the c suggested by other authors, to discover the conditions under which it was better (i.e. having lesser criteria) than both c(min), c(σ) on criterion (ii), (i) respectively. In the large sample study, variances and covariances were found (in many cases approximately) for all estimates of the parameters entering into the conditional probabilities of misclassification (level (c)). Extension of results to level (b) and to special cases of level (a) were given. From these variances and covariances were derived the expectations of these probabilities for both criteria, at level (c), and comparisons were made where feasible. Results were tabulated. / Ph. D.

The personal computer furnishes engineers and managers with an appropriate tool for analysis of statistical problems particularly in areas such as quality control. Following a review of the literature, no simple, low cost software packages were found that allow the personal computer user to design, execute, and statistically analyze physical processes while providing graphical display of the results. The primary objective of this research is to develop a software system which runs on a personal computer and provides management with an affordable vehicle for statistical analysis of problems, particularly those faced by the decision maker. This software system includes the design of inferential procedures as well as their execution. Many of the inferential procedures are based on the assumptions of independent and normally distributed observations, and in the case of confidence intervals and hypothesis tests for means, offers a technique which assists the user in overcoming problems associated with correlated sample data. The software system attempts to detect when these assumptions appear to be violated. In the case of independence, a methodology is applied which attempts to offset the deleterious effects of the correlation on parameter estimates. Although statistical methods are used to interpret sampled data, extensive knowledge of statistics is not necessary to use this software. The software is intended for managers and quality engineers who possess knowledge of basic statistics, the ability to interpret control chart information, and a basic working knowledge of personal computers. Execution requires an IBM PC (or a true compatible) which contains a color graphics card and at least 512K of memory. Both color and monochrome monitors are supported. The software is written in Turbo Basic 1.0, a product of Borland International. The software operates under the MS-DOS 2.0 (or later version) operating system. DOS, Disk Operating System, provides for communication between the keyboard, the disk drives, and the system unit. / Master of Science