Congratulations 2004-2005 Graduates
Dr. Hongmei Jiang performed her Ph.D. research in the area of statistical bioinformatics under the direction of Professor Rebecca Doerge. The title of Hongmei's dissertation is: "A two-step procedure for multiple pairwise comparisons in microarray experiments."
Two major (and novel) statistical contributions resulted from Dr. Jiang's work. The first piece pertains to extending Hochberg and Benjaminis (1995) false discovery rate (FDR) to include an estimate of the proportion of true null hypotheses. Existing methods for estimating this proportion yield either large bias or large variance. Hongmei developed a simple and easy to implement method to estimate the proportion of true null hypotheses that has both small bias and small variance. The combination of Hongmei's estimate with Benjamini and Hochbergs FDR controlling procedure controls the FDR below, but extremely close to alpha. This improved approach is then implemented in the second major result from Dr. Jiang's research, a two-stage testing procedure for comparing several treatments. The purpose and motivation of this two-stage approach is to improve the power to identify differential expression for all pairwise comparisons when there are a large number of genes to be tested. Because the FDR controlling procedure is applied to a large family of pairwise comparisons, the impact is a reduction of power. To address this problem, and to improve the power of detecting differentially expressed genes, Hongmei proposed a novel two-step multiple comparison procedure. The greatest challenge in doing this is to choose the significance levels in the respective steps of the two-step procedure so that the overall FDR of the two-step procedure is controlled at a pre-specified significance level. Hongmei put forth a set of recommendations on how to choose the significance levels at each step of the two-step procedure. In turn, Hongmei's two-step procedure has more power than the typical one-step procedure in terms of detecting significant differential expression because the number of pairwise comparisons is greatly reduced in the second step.
Dr. Jiang will start as an Assistant Professor in the Department of Statistics at Northwestern University, September 2005.
Dr. Chuancai Wang performed his Ph.D. research in the area of statistical bioinformatics under the direction of Professors Bruce A. Craig and Jun Xie. The title of his dissertation was: "Motif Discovery via Context Dependent Models."
Dr. Wang focused his research on the discovery of transcription factor binding sites. Identification of these motifs in the genome is vital for the understanding of gene regulation and function. Many of the current algorithms/models available to biologists make strong, often oversimplifying assumptions regarding the structure of the DNA background and motif sequences. Dr. Wang developed a more flexible motif algorithm/model that has shown great promise for finding transcription factor binding sites and will likely be adapted to protein motif detection.
Dr. Wang will start as an Assistant Professor in the Division of Biostatistics (Department of Health Evaluation Sciences), Penn State College of Medicine in July 2005.
Dr. Stevens performed his Ph.D. research in the area of statistical bioinformatics under the direction of Professor Rebecca Doerge. The title of John's dissertation is: "Meta-analytic approaches for microarray data".
John developed both a frequentist and Bayesian meta-analysis approach for the analysis of Affymetrix microarray data. He proposed a statistically based meta-analytic approach to microarray analysis for the purpose of systematically combining results from the different laboratories. This approach provides a more precise view of genes that are significantly related to the condition of interest while simultaneously allowing for differences between laboratories. A simulation model based on the Affymetrix platform examined the adaptive nature of the meta-analytic approach and illustrated the usefulness of such an approach in combining microarray results across laboratories. Dr. Stevens' proposed meta-analysis has been applied to a series of mouse data experiments across multiple laboratories that are each independently studying multiple sclerosis in humans.
Dr. Stevens will start as an Assistant Professor the Department of Mathematics and Statistics, Utah State University, June 2005.
Dr. Olga Vitek (May '05) performed her Ph.D. research in the area of statistical bioinformatics under the direction of Professors Bruce A. Craig and Chris Bailey-Kellogg. The title of her dissertation was: "An Inferential Approach to Protein Backbone Nuclear Magnetic Resonance Assignment."
Dr. Vitek focused her research on statistical issues associated with the assignment and inference of the protein backbone using nuclear magnetic resonances (NMR). Nuclear magnetic resonance spectroscopy is a key experimental technique used to study structure, dynamics and molecular interactions of proteins. NMR methods face the bottleneck of spectral analysis, in particular determining the resonance assignment, which helps define the mapping between atoms in the protein and peaks in the spectra. A substantial amount of noise in spectral data, along with ambiguities in interpretation, make this analysis a daunting task, and there exists no generally accepted measure of uncertainty associated with the resulting solutions. Dr. Vitek developed a Bayesian statistical model that accounts for various sources of uncertainty and provides an automatable framework for inference. While assignment has previously been viewed as a deterministic optimization problem, she demonstrated the importance of considering all solutions consistent with the data, and developed an algorithm to search this space within her statistical framework.
Dr. Vitek currently has a post-doctorial position with Dr. Ruedi Aebersold in the Institute for Systems Biology at the University of Washington. She will start as an Assistant Professor in the Department of Statistics at Purdue University in August 2006.
Past graduates can be seen here.