Chapter 5 194 Supplemental material Supplementary results 1. Development of the test The following pages provide heatmaps for samples in the development cohort of the features used for each of classifier A, classifier B, classifier C, and for the test. Samples are sorted into the classification groups of each classifier or the test. The mass spectral features are labelled on the x-axis. The heatmap plots log10 of the feature values, with the red/brown for the largest features and dark blue for the smallest features. 2. Reproducibility of Test Classifications To assess the reproducibility of the test, the test was run from scratch on samples from validation set 1 on two occasions more than one year apart. The classifications obtained for run1 and run2 are compared in Supplementary Table S9 for classification sensitive vs. intermediate vs. resistant, in Supplementary Table S10 for the binary combination resistant vs. not resistant (intermediate and sensitive), and in Supplementary Table S11 for the binary combination not sensitive (intermediate and resistant) vs. sensitive. Classification concordance is 85% for the three-way classifications, 91% for the resistant / not resistant combination and 93% for the not sensitive / sensitive combination.
RkJQdWJsaXNoZXIy MTk4NDMw