Malignant gliomas are high‑grade gliomas, which are derived from glial cells in the spine or brain. To examine the mechanisms underlying malignant gliomas in the present study, the expression profile of GSE54004, which included 12 grade II astrocytomas, 33 grade III astrocytomas and 98 grade IV astrocytomas, was downloaded from the Gene Expression Omnibus. Using the Limma package in R, the differentially expressed genes (DEGs) in grade III, vs. grade II astrocytoma, grade IV, vs. grade II astrocytoma, and grade IV, vs. grade III astrocytoma were analyzed. Venn diagram analysis and enrichment analyses were performed separately for the DEGs using VennPlex software and the Database for Annotation, Visualization and Integrated Discovery. Protein‑protein interaction (PPI) networks were visualized using Cytoscape software, and subsequent module analysis of the PPI networks was performed using the ClusterONE tool. Finally, glioma‑associated genes and glioma marker genes among the DEGs were identified using the CTD database. A total of 27, 1,446 and 776 DEGs were screened for the grade III, vs. grade II, grade IV, vs. grade II, and grade IV, vs. grade III astrocytoma comparison groups, respectively. Functional enrichment analyses showed that matrix metalloproteinase 9 (MMP9) and chitinase 3‑like 1 (CHI3L1) were enriched in the extracellular matrix and extracellular matrix structural constituent, respectively. In the PPI networks, annexin A1 (ANXA1) had a higher degree and MMP9 had interactions with vascular endothelial growth factor A (VEGFA). There were 10 common glioma marker genes between the grade IV, vs. grade II and the grade IV, vs. grade III comparison groups, including MMP9, CHI3L1, VEGFA and S100 calcium binding protein A4 (S100A4). This suggested that MMP9, CHI3L1, VEGFA, S100A4 and ANXA1 may be involved in the progression of malignant gliomas.