The key molecular changes in the progression of glioma are closely related to tumor heterogeneity, pathological grade, precision treatment and prognosis of glioma. At present, a visually quantitative assessment criteria about the key molecular typing of glioma is still absent. Based on the previous research, this project intends to establish a multi-dimensional database of glioma from clinical, radiomics and microomics levels. The investigators aim to filter out the specific molecular markers in the progression of glioma and explore the intrinsic connection of radiomics features and microomics molecular markers by using bioinformatics integration analysis and artificial intelligence multiple kernel learning. Thus, the investigators could determine the specific molecular mechanism in the progression of glioma, and establish a visually quantitative assessment system of pre-operative precisive grading, molecular typing discrimination and prognosis prediction. The completion of this project is of great significance for improving molecular diagnostic level of glioma, guiding individualized diagnosis and treatment decisions, and improving the survival rate of patients.