Anaplastic thyroid carcinoma (ATC) is a rare but extremely lethal malignancy. However, little is known about the pathogenesis of ATC. Given its high mortality, it is critical to improve our understanding of ATC pathogenesis and to find new diagnostic biomarkers. In the present study, two gene microarray profiles (GSE53072 and GSE65144), which included 17 ATC and 17 adjacent non-tumorous tissues, were obtained. Bioinformatic analyses were then performed. Immunohistochemistry (IHC) and receiver operating characteristic (ROC) curves were then used to detect transmembrane protein 158 (TMEM158) expression and to assess diagnostic sensitivity. A total of 372 differentially expressed genes (DEGs) were identified. Through protein-protein interaction (PPI) analysis, we identified a significant module with 37 upregulated genes. Most of the genes in this module were related to cell-cycle processes. After co-expression analysis, 132 hub genes were selected for further study. Nine genes were identified as both DEGs and genes of interest in the weighted gene co-expression network analysis (WGCNA). IHC and ROC curves confirmed that TMEM158 was overexpressed in ATC tissue as compared with other types of thyroid cancer and normal tissue samples. We identified 8 KEGG pathways that were associated with high expression of TMEM158, including aminoacyl-tRNA biosynthesis and DNA replication. Our results suggest that TMEM158 may be a potential oncogene and serve as a diagnostic indicator for ATC.