Submitted to: Applied Statistics In Agriculture Conference Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 1/22/2007
Publication Date: 4/30/2007
Citation: Boykin, D.L., Taliercio, E.W., Kelley, R.Y., Williams, W.P. 2007. A Visual Aid for Statisticians and Molecular Biologists Working With Microarray Experiments. Proceedings of the Eighteenth Annual Kansas State University Conference on Applied Statistics in Agriculture. p.33-49. Interpretive Summary:
Technical Abstract: The use of microarrays to measure the expression of large numbers of genes simultaneously is increasing in agriculture. Statisticians are expected to help biologists analyze these large data sets to identify biologically important genes that are differentially regulated in the samples under investigation; however molecular biologists are often unfamiliar with the statistical methods used to analyze microarrays. Methods are developed to graphically represent microarray data and various types of errors commonly associated with microarrays to help visualize sources of error. Two case studies are used. In case study one, genes differentially regulated when two corn lines, one resistant and one sensitive to aflatoxin, are treated with Aspergillus flavus isolate NRRL 3357 or left untreated are investigated. Analyses and images showing 3 types of variation are shown. Genes are ranked according to fold change and re-ranked after adjusting for potential sources of error. In case two, genes differentially regulated in 1-day-old fiber compared to whole ovules or older fibers are investigated. Data and sources of error are imaged as described for case one and genes with significant changes in gene expression are identified.