Background: Tongue cancer, as one of the most malignant oral cancers, is highly invasive and has a high risk of recurrence. At present, tongue cancer is not obvious and easy to miss the opportunity for early diagnosis when in the advanced stage. It is important to find markers that can predict the occurrence and progression of tongue cancer.
Methods: Bioinformatics analysis plays an important role in the acquisition of marker genes. GEO and TCGA data are very important public databases. In addition to expression data, the TCGA database also contains corresponding clinical data. In this study, we screened three GEO data sets that met the standard, which included GSE13601, GSE34105, and GSE34106. These data sets were combined using the SVA package to prepare the data for differential expression analysis, and then the limma package was used to set the standard to p < 0.05 and |log2 (FC)| ≥ 1.5.
Results: A total of 170 differentially expressed genes (DEGs) were identified. In addition, the DEseq package was used for differential expression analysis using the same criteria for samples in the TCGA database. It ended up with 1,589 DEGs (644 upregulated, 945 downregulated). By merging these two sets of DEGs, 5 common upregulated DEGs (CCL20, SCG5, SPP1, KRT75, and FOLR3) and 15 common downregulated DEGs were obtained.
Conclusions: Further functional analysis of the DEGs showed that CCL20, SCG5, and SPP1 are closely related to prognosis and may be a therapeutic target of TSCC.
Keywords: CCL20; DEGs; SCG5; SPP1; tongue cancer.
© 2023 Lv and Yu.