An integrated analysis of cancer genes in thyroid cancer

Cancer driver genes are commonly mutationally disrupted in cancer, which confers a growth advantage to tumor cells. Recent studies preferentially search for recurrently mutated driver genes across multiple tumor samples, leading to the neglect of low-frequency mutated cancer genes. The present study was conducted to identify cancer‑driving genes in thyroid cancer with two distinct tools, OncodriveFM and Dendrix, which aim to detect neglected driver genes with low mutation frequency. A total of 23,620 somatic mutations generated by whole‑exome sequencing of 446 tumor/normal pairs of thyroid cancer were obtained from TCGA. Variant classification was conducted with Ensembl Variant Effect Predictor (VEP). OncodriveFM and Dendrix were applied to detect driver genes and pathways with statistical evidence. In addition, we analyzed DNA‑methylation status, copy number variation, expression levels and fusion genes among these driver candidates. In total, non‑synonymous mutations accounted for over 55% (13,091/23,620) of the total variants; 53 and 3 driver genes were determined by OncodriveFM and Dendrix, respectively, including 6 recurrently mutated driver genes, such as BRAF, NRAS, HRAS, EIF1AX, KRAS and 47 new genes. A total of 75 pathways with high function impact bias were identified by OncodriveFM. Two genes, FHOD3 and SRP72, were hypomethylated, overexpressed and involved in major deletions in thyroid cancer. Moreover, we identified 91 pairs of fusion genes, 89 of which were new fusion pairs in thyroid cancer. In conclusion, we successfully identified a list of new cancer genes, pathways and fusion genes, providing better insight into the tumorigenesis of thyroid cancer.