Statistical Perspectives in Some Genomic and Environmental Studies

Abstract

Statistical Perspectives in Some Genomic and Environmental Studies Pranab Kumar Sen Dept. of Statistics and Operations Research, Univ. of North Carolina at Chapel Hills, USA ?: The presentation aims to focus on the primary role of nonstandard statistical methodology in some recent interdisciplinary field with emphasis on some environmental and genomic studies. Specifically, the following problems are used as illustration: 　　(i) Ongoing Prostate Cancer Study in Finland (Fischer et al. ); 　　(ii) Gingivitis (Dental) Biomarker Study at Chapel Hill (Preisser et al. 2011); 　　(iii) Congener profile environmental toxicity study in Taiwan (Chen et al. 2012); and 　　(iv) SARS RNA study in China, Hong Kong, Taiwan and Thailand. In all these studies, standard parametric procedures may not work well. Further, there is a persistent high dimension low sample size (HDLSS) undercurrent in all these data models, and the different variables (eg., gene expressions) are not independent. Nonparametric HDLSS tools are incorporated to validate robust statistical procedures. Open problems are also discussed.