| 總結: | We aimed to identify potential molecular biomarkers through integrated analysis of microbiome and transcriptome data in colon cancer. Microbiome, transcriptome, and clinical phenotype data were downloaded from TCMA and TCGA databases. Intestinal flora subtypes were identified, followed by screening feature genes of each subtype. Univariate and LASSO Cox regression analyses were conducted to identify optimal prognostic genes to establish prognostic model. Survival, clinical phenotypes, pathways, and immune infiltration within different risk groups were compared. Three intestinal flora subtypes were identified: Proteobacteria, Fusobacteria, and Bacteroidetes subtypes. Bacteroidetes subtype tended to have worse survival. Feature genes of Bacteroidetes subtype were mainly implicated in immune-related functional terms, whereas those of Fusobacteria and Proteobacteria subtypes were mainly implicated in functions associated with inflammatory response and transcription, respectively. LASSO Cox regression identified 6 optimal prognostic genes: CSF1R, HLA-DOA, NOS3, HOXB4, PLA1A, and RPL3, to establish prognostic risk model. High-risk patients were associated with worse survival and had high proportions of Bacteroidetes subtype and high immune infiltration. Proteobacteria, Fusobacteria, and Bacteroidetes largely dominated the gut microbial flora in colon cancer patients. Prognostic model based on feature genes of intestinal flora subtypes can predict survival in colon cancer.
|
|---|