Fig. 2

Validation of tree classifiers on novel datasets. A Time and oxygen series dataset. Each of the dots represents one of the samples in the five time series in PRJNA561635, ordered by oxygen tension and time. The color of each dot represents the mean probability of each sample to be classified as hypoxic, while the size of the dot is proportional to the number of trees correctly classifying each sample. B–D Distribution of the accuracy of the 276 classification trees on validation datasets. Different colors are used to identify trees including specific genes and/or topology as indicated in the legend. See text for details. B 22 datasets from RNA fractions other than total mRNA. C 2 datasets of clear cell renal carcinoma with VHL mutations and paired adjacent healthy tissues. D 34 mouse RNA-seq datasets