Fig. 7

Validation of classifier in an independent dataset. The ability of the host-based classifier to predict outcome was tested in an independent dataset where the Ct values all lay between 20 and 22. a ROC showing the host-based classifier (blue) in the independent dataset compared to the Ct value (green). The line represents the line y = x which shows where the ROC falls if the predictions are equal to random selection. b pROC showing the comparison for the host classifier (blue) and Ct value (green) with a false positive rate up to 0.2. In the validation dataset, the host classifier was able to predict better than the Ct value