Identification of prognostic miRNA biomarkers for predicting overall survival of colon adenocarcinoma and bioinformatics analysis: A study based on The Cancer Genome Atlas database

Object: This study was conducted to identify the prognostic microRNA (miRNA) for the prediction of survival in colon adenocarcinoma (CA).

Methods: miRNA profiling of patients with CA was downloaded from The Cancer Genome Atlas (TCGA) database. After data processing, univariate Cox regression was performed to select potential prognostic miRNAs. Least absolute shrinkage and selection operator approach was conducted to identify the key prognostic miRNA biomarkers. Log-rank test was performed to compare survival outcome between patients with different regulation type of identified miRNAs. Then, bioinformatics analysis including Gene Ontology, Disease Ontology, and pathway enrichment analysis was performed on the selected prognostic miRNAs. A nomogram was generated based on the multivariate Cox proportional hazard model to illustrate the association between the identified miRNAs and the survival of patients with colorectal cancer (CRC). All analyses were conducted with packages in the R software.

Results: A total of 1881 miRNAs were obtained from TCGA database in which 15 miRNAs were finally selected in multianalysis with covariates and incorporated into functional annotation analyses. Log-rank test suggested that the identified miRNAs were associated with the survival of CA. Seven out of 15 selected miRNAs were identified for the first time in CA. Bioinformatics analysis suggested that the identified miRNAs were associated with the development and progress of CRC. The association between the identified miRNAs and the survival of patients with CRC were presented in a nomogram.

Conclusion: The seven newly identified miRNAs might be potential prognostic biomarkers for the survival of CA. Further research about the selected miRNAs were worth to conduct.