Fig. 2

The four validated downregulated miRNAs may serve as significant prognostic markers in CRC classification. a An ROC curve built on a univariate classification model based on miR-195 expression across independent TCGA dataset for predicting CRC. b Performance of the four miRNAs on CRC tissue classification. CRC risk score was built using a linear regression model by R software. c Kaplan–Meier plots for overall survival for a discriminatory median miR-195 expression, from TCGA sequencing data to assess prognostic accuracy. P values were calculated using the log-rank test. d Forest plots summarizing the downregulation of hsa-miR-195 from eight datasets in the integrated analysis. Each row represents a study with standardized mean difference between CRC and normal tissues (black dot) and the confidence interval of 95% (dark line). The size of the gray box is proportional to the relative effect size of each dataset. The dotted vertical line at 0 represents the null hypothesis. The diamonds represent overall, combined mean difference for miR-195. Thus, negative values indicate downregulation of miR-195 in CRC. e Venn diagrams of putative miR-195 targets predicted by TargetScan (v7.1), miRDB, PicTar, and DIANA-TarBase (v7.0). f Top ten KEGG pathways that are enriched for the miR-195 targets are mainly cancer-specific pathways