File S1 - Detection of Patient Subgroups with Differential Expression in Omics Data: A Comprehensive Comparison of Univariate Measures

Figure S1, In-depth comparison of different methods' AUC values for the composite null situation. The study includes all combinations of sample sizes , subgroup proportions , and scenarios with . AUC values are plotted against the deviation . Figure S2, In-depth comparison of different methods' AUC values for the simple null situation. The study includes all combinations of sample sizes , subgroup proportions , and scenarios with . AUC values are plotted against the deviation . Figure S3, (log2-)intensity plots of the top candidates for ParkCHIP data. To compare the considered methods, the respective top 15 candidates of each method are shown in (log2-)intensity plots. We used controls and Parkinson patients from the ParkCHIP project. Table S1, Overview scenario I, composite null situation. Best tests for all combinations of sample size and subgroup proportions in scenario I, , neatly arranged in a color-coded table. Table S2, Overview scenario II, composite null situation. Best tests for all combinations of sample size and subgroup proportions in scenario II, , neatly arranged in a color-coded table. Table S3, Overview scenario III, composite null situation. Best tests for all combinations of sample size and subgroup proportions in scenario III, , neatly arranged in a color-coded table. Table S4, Overview scenario I, simple null situation. Best tests for all combinations of sample size and subgroup proportions in scenario I, , neatly arranged in a single table. Table S5, Overview scenario II, simple null situation. Best tests for all combinations of sample size and subgroup proportions in scenario II, , neatly arranged in a single table. Table S6,Overview scenario III, simple null situation. Best tests for all combinations of sample size and subgroup proportions in scenario III, , neatly arranged in a single table.
(PDF)