Fig. 3.

Signature classification performance by age. A Weighted mean AUCs were generated for each signature’s classification of bacterial patients and viral patients across pediatric-only (red) and adult-only (blue) datasets. Values were weighted based on the number of subjects in a validation dataset. The distributions of such weighted mean AUCs were plotted, and significance was determined by the Wilcoxon rank-sum test. B After pooling samples across datasets, each signature’s accuracy was calculated and plotted for five age groups (<3 months, 3 months–1 year, 2–11 years, 12–18 years, and adult). This plot shows the median and IQR of each signature’s accuracy in each age group for bacterial and viral classification. * indicates p < 0.05 as compared to the adult population