Numerous colon cancer cases are resistant to chemotherapy based on oxaliplatin and suffer from relapse. A number of survival- and prognosis-related biomarkers have been identified based on database mining for patients who develop drug resistance, but the single individual gene biomarker cannot attain high specificity and sensitivity in prognosis prediction. This work was conducted aiming to establish a new gene signature using oxaliplatin resistance-related genes to predict the prognosis for colon cancer. To this end, we downloaded gene expression profile data of cell lines that are resistant and not resistant to oxaliplatin from the Gene Expression Omnibus (GEO) database. Altogether, 495 oxaliplatin resistance-related genes were searched by weighted gene co-expression network analysis (WGCNA) and differential expression analysis. As suggested by functional analysis, the above genes were mostly enriched into cell adhesion and immune processes. Besides, a signature was built based on four oxaliplatin resistance-related genes selected from the training set to predict the overall survival (OS) by stepwise regression and least absolute shrinkage and selection operator (LASSO) Cox analysis. Relative to the low risk score group, the high risk score group had dismal OS (P < 0.0001). Moreover, the area under the curve (AUC) value regarding the 5-year OS was 0.72, indicating that the risk score was accurate in the prediction of OS for colon cancer patients (AUC >0.7). Additionally, multivariate Cox regression suggested that the signature constructed based on four oxaliplatin resistance-related genes predicted the prognosis for colon cancer cases [hazard ratio (HR), 2.77; 95% CI, 2.03–3.78; P < 0.001]. Finally, external test sets were utilized to further validate the stability and accuracy of oxaliplatin resistance-related gene signature for prognosis of colon cancer patients. To sum up, this study establishes a signature based on four oxaliplatin resistance-related genes for predicting the survival of colon cancer patients, which sheds more light on the mechanisms of oxaliplatin resistance and helps identify colon cancer cases with a dismal prognostic outcome.

The morbidity and mortality of colorectal cancer (CRC) remained to be very high worldwide. Recently, circRNAs had been revealed to have a crucial role in cancer prognosis and progression. Numerous researches have shown that RNA sequencing technology and in silico method were widely used to identify pathogenic mechanisms and uncover promising targets for diagnosis and therapy. In this study, these methods were analyzed to obtain differentially expressed circRNAs (DECs). We identified upregulated 316 circRNAs and reduced 76 circRNAs in CRC samples, in comparison with those in normal tissues. In addition, a competitive endogenous network of circRNA-miRNA-mRNA was established to predict the mechanisms of circRNAs. Bioinformatics analysis revealed that these circRNAs participated in metabolism regulation and cell cycle progression. Of note, we observed the hub genes and miRNAs in this ceRNA network were associated with the survival time in CRC. We think this study could provide potential prognostic biomarkers and targets for CRC.