Figure 6 | Scientific Reports

Figure 6

From: An algorithm-based meta-analysis of genome- and proteome-wide data identifies a combination of potential plasma biomarkers for colorectal cancer

Figure 6

Classification accuracy of different combinations of two, three and four proteins in a test set of 8 patients and control samples. Boxplot presenting discriminatory function score based on the protein expression of combinations of two, three and four proteins in plasma samples of patients and controls in test group measured with ELISA. The protein combinations were: (A) TRIM28 and PLOD1, (B) TRIM28, PLOD1 and CEACAM5, C) TRIM28, PLOD1, CEACAM5 and P4HA1. Discriminatory function score was calculated as the sum of the concentrations of respective proteins. The discriminatory cut-off value is presented with the solid purple vertical line. The dark and bright orange dots denote values obtained for patients and controls in the test group respectively. The bars in the boxes represent median, 25th and 75th percentiles, while whiskers extend to ±2.7 σ. For the combination of two proteins we obtained best results when summing concentrations of TRIM28 and PLOD1 (sensitivity of 88%, specificity of 75%). For the combination of three proteins we yielded sensitivity of 100% and specificity of 88% (sum of TRIM28, PLOD1, and CEACAM5). However, best separation was possible when using four proteins, namely TRIM28, PLOD1, CEACAM5, and P4HA1 (both sensitivity and specificity of 100%).

Back to article page