The processes of oncogenesis are multiple steps, which include the activation of oncogenes, inactivation of tumor suppressor genes, and the mutations of DNA repair genes and other cancer-related factors. All the above phenomena cause changes in the processes of cell growth, differentiation, reproducing, and apoptosis. Especially, the over-expression of oncogenes plays a very important role in the initial state of numerous cancer cases. On the other hand, tumor suppressor genes are the brakes of growth. Loss of function in tumor suppressor genes makes the tumor formation more possible. Many tumor suppressor genes had been identified in the past. However, oncogenes that have been identified present only the tip of the iceberg. More oncogenes are waiting us to discover. Mining the human genome to identify genetic mutations that cause cancer is like looking for needles in a haystack, especially by using the traditional one-by-one analysis method. For now, cDNA microarray can be used to analyze the changes of the gene expression in different cancer cells in a high throughput manner. A lot of microarray data had been accumulated from the cancer-related researches in the past. It is quite interesting to mine useful information from these data to find possible cancer-related genes. In this thesis, we propose a bioinformatic approach to discover possible cancer-related genes based on differential expression of cDNA microarray databases for cancer and normal tissues. The t-test is also used to assess the significance of the differential expression. The proposed cancer-related genes discovering system will narrow down the range of bio-experiments and avoid some unnecessary experiments. It can really help the biologist to seek the cancer-related genes more efficient.
