医学统计学（扬州大学）Introduction to Statistical Analysis of Gene Expression Data.pptVIP

Introduction to Statistical Analysis of Gene Expression Data Feng Hong Beespace meeting April 20, 2005 The Central Dogma DNA Transcription RNA Translation Protein Source: / A gene is a sequence of nucleotides that codes for a protein All cells contain the same gene information in DNA, but only a few genes are expressed in certain cell The presence of mRNA in a cell indicates that a gene is active; Microarray Technololgy /RC/VL/GG/microArray.html Microarray Examine how active the thousands of genes are at once Florescent-dye-labeled mRNA from different samples hybridize to the DNA on the array Intensity of florescent indicates the expression level of the gene in the sample Steps in Microarray experiment Experimental Design Signal Extraction Image Analysis Normalization: remove the artifacts across arrays Data Analysis Selection of Genes differentially expressed Clustering and classification Experimental Design For two-color cDNA experiment, only two sample mRNA can be hybridized on the one array Factors influencing choice of experimental design Number of different samples Aim of the experiment: which comparisons are of primary interest Constraint of resources Power of the experiment Experimental Design Direct Comparison : compare only two mRNA samples Dye-swap is recommended to minimize the Reference Sample: Compare several samples with reference Indirect comparison between the samples Saturated Design More than two MRNA samples All comparison are of interest Loop Design Used in time couse More complicated designs Design used in Whitfield et al.(2003) Source: Whitfield, Cziko, Robinson, 2003, Gene Expression Profiles in the brain predict behavior in individual honey bees, Science, supplement materials Gene expression measurements Gene expression data are noisy Source of errors Microarray manufacturing Preparation of mRNA from biological samples Hybridization Scanning Imaging Image Analysis Preprocess the raw scanned image Grid