Gene expression profiling facilitates the understanding of biological characteristics of gliomas. Previousstudies mainly used regression/variance analysis without considering various background biological andenvironmental factors. The aim of this study was to investigate gene expression differences between grade IIIand IV gliomas through partial least squares (PLS) based analysis. The expression data set was from the GeneExpression Omnibus database. PLS based analysis was performed with the R statistical software. A total of1,378 differentially expressed genes were identified. Survival analysis identified four pathways, including Priondiseases, colorectal cancer, CAMs, and PI3K-Akt signaling, which may be related with the prognosis of thepatients. Network analysis identified two hub genes, ELAVL1 and FN1, which have been reported to be relatedwith glioma previously. Our results provide new understanding of glioma pathogenesis and prognosis with thehope to offer theoretical support for future therapeutic studies.