Reliability Verification of Microarray Gene Expression Database

[[abstract]]Abstract Cancer is due to gene mutation. At present, there are two kinds of genes, oncogene and tumor suppressor gene, involved in cancer. Cancer is usually resulted from the activation of oncogenes, malfunction of tumor suppressor genes, and mutation of DNA repair genes or other cancer-related factors. cDNA microarray can be used to analyze the gene expression in different cancer cells. Enormous researches have been studied in this area. When analyzing microarray data using T-test, the α-value is usually set to 0.05 no matter what kind of data is considered? The purpose of this thesis is to discuss whether the value of α = 0.05 is valid or not in analyzing various microarray gene expression data. We attempt to find out a more precise α value to identify gene markers. In our study, we have collected microarray gene expression data of liver and breast cancers. We first sieve out uncertain data. Then, we analyze the reliability of microarray gene expression databases using T-test. Finally, we choose a reliable α value according to ratios of over and under expression values. We found the α value must be set according to different microarray experiments. We hope the result is useful for the biologists to find out information and to verify cancer-related genes effectively