Statistical Issues in cDNA Microarray Data Analysis

摘要: Statistical considerations are frequently to the fore in analysis of microarray data, as researchers sift through massive amounts data and adjust for various sources variability order identify important genes amongst many which measured. This article summarizes some issues involved provides a brief review tools available deal with them. Any experiment involves number distinct stages. Firstly there is design experiment. The must decide be printed on arrays, RNA hybridized arrays how hybridizations will replicated. Secondly, after hybridization, follows data-cleaning steps or `low-level analysis’ data. images processed acquire red green foreground background intensities each spot. acquired red/green ratios normalized dye-bias any systematic variation other than that due differences between samples being studied. Thirdly, analyzed by graphical numerical means select differentially expressed (DE) find groups whose expression profiles can reliably classify different into meaningful groups. sections this correspond roughly steps. following notation used throughout article.