Data Mining Of Fusion Genes In Human Cancer RNA-seq Data And Database Construction

Cancer is a chronic disease of mankind that has not been overcome yet. Its cause is very complex. In addition to genetic factors in patients, environments, habits and contact with carcinogens, other factors are also related to carcinogenesis. The occurrence of cancer at the molecular level appears as acquired gain-of-function mutations in oncogenes and loss-of-function mutations in tumor suppressor genes, which will lead to abnormal cell proliferation. When developing to a certain stage, it will invade surrounding normal tissues and even spread to other parts of the body through the blood or lymphatic system. Currently more than a hundred kinds of cancers have been found in humans, and according to epidemiological studies, cancer is still highly prevalent and the global morbidity and mortality of cancers are higher than that of other diseases. However, the incidence and prognosis vary greatly in different cancers. Since patients with the same kind of cancer may also have different etiology, and the genetic states of cells within the single tumor site of the a patient are still of great difference. These all make the pathogenesis study and treatment of cancer very difficult. Studies have shown that genetic material changes are the causes of cancer, including single nucleotide changes, indels of small fragments and the chromosome structure changes. As a result of chromosome insertions, deletions or inversions which cause abnormal chromosome structure, fusion genes have been proven to be one of the possible causes of cancer, such as BCR-ABL1 in chronic myeloid leukemia. However, relying solely on fluorescence in situ hybridization and other experimental techniques for fusion gene mining obviously does not meet the heterogeneity of cancer. Thanks to the rapid development of bioinformatics analysis technology and second-generation sequencing technology in recent years, they have been more and more widely used to detect fusion genes because of their high precision, high flux as well as the persistently decreasing cost, and greatly promot the gene fusion mining in cancer genome/transcriptome data. Therefore in this opportunity, we make use of a variety of human cancer transcriptome sequencing data that have been openly published to excavate fusion genes. After performing necessary quality control and filtering steps, we conducted an in-depth fusion gene detection in colorectal cancer study by analyzing the functions of genes involved in the fusion events and the impacts of gene structure changes. We looked for fusion genes which more likely playing a driving role in the development of cancer, and tried to explain the possible pathogenic molecular mechanism of these fusion genes. At the same time, a database of fusion genes only applying second generation sequencing data was constructed, in which fusion genes with relatively high prevalence in each cancer type were listed and provided a potential target for cancer research.