Fig. 6

Classification matrices generated by combining the complementary classification strengths of different proteins. All possible combinations of 1 to 6 proteins were used for classification. The classification matrix for the best combination of A 6 proteins, B 4 proteins, C 3 proteins, and D 2 proteins. In A the probabilities enclosed in blue represent false negative classifications of tumor educated platelets (TEPs). Summing the probabilities enclosed in blue together, the average probability for false negative classification of TEPs, defined as the average probability to classify a platelet of category EFO21, MCF7 or 231 into the categories R or 10A is less than 6%. Similarly, the probabilities enclosed in red represent false positive classifications (with a probability < 8% when summed together), and those in black true positive classifications of TEPs (> 90% when summed together)