Fig. 2

Construction of three-marker diagnostic model. A The sensitivity and specificity of selected 12 candidate markers in 9 UC cell lines and 12 healthy people urine samples. Testing method, qMSP. B The circle bar plot showing the respective positive cases of 6 candidate markers in UC patients (N = 20), non-neoplastic individuals (N = 20) and other tumor patients (N = 20). Testing method, qMSP. C The pie chart showing the accuracy of 6 candidate markers (corresponding y-axis value of each pie). The components of each pie represent the numbers of true positive case in UC group (red), true negative case in non-neoplastic group (blue) and true negative case in other tumors group (purple). The x-axis represents gene names of 6 candidate markers. D The ROC curves of four diagnostic markers, accuracy top 3 genes (VIM, TMEM220 and PPM1 N) and Combine Parallel (three-marker diagnostic model), shown in the graph for distinguishing 20 UC patients from other 40 non-UC cases. The AUC values for each diagnostic markers were marked in the graph legend. Three-marker diagnostic model was based on the logic of single-gene positive indicating positive and triple negative indicating negative