Molecular Classification Based on Prognostic and Cell Cycle-Associated Genes in Patients With Colon Cancer

The molecular classification for colon cancer patients is inconclusive, the gene set enrichment analysis (GSEA) of dysregulated genes among normal and tumor tissues indicated cell cycle played a crucial role in colon cancer. We performed univariate COX regression analysis to find out the prognostic-related genes, and these genes were then intersected with cell cycle-associated genes and were further recognized as prognostic and cell cycle-associated genes. Unsupervised non-negative matrix factorization (NMF) clustering was performed based on cell cycle-associated genes. 2 subgroups were identified with different overall survival, clinical features, cell cycle enrichment profile and mutation profile. Through nearest template prediction (NTP), the molecular classification could be effectively repeated in the original dataset and validated in several independent datasets indicating the classification highly repeatable. Furthermore, we constructed 2 prognostic signatures in 2 subgroups respectively. Our molecular classification based on cell cycle may provide novel insight into the treatment and the estimation of prognosis of colon cancer.