Fig. 4

This figure displays the sensitivity and specificity across the 9 different simulation settings accounting for different numbers of average cells per subject and different levels of technical noise. The points represent the median sensitivity/specificity and the bars represent the first and third quartiles across 50 iterations for each simulation setting. (A)-(C) display the sensitivity for \(\boldsymbol {\mathcal {G}}\) using MOSCATO, AUC selections using Bonferroni adjusted p-values, and AUC selections using cutoffs (<0.3 or >0.7). (D)-(F) display the specificity for \(\boldsymbol {\mathcal {G}}\) under the three methods in the same order. Similarly, (G)-(I) displays the sensitivity for \(\boldsymbol {\mathcal {X}}\) and (J)-(L) displays the specificity for \(\boldsymbol {\mathcal {X}}\). Under perfect selections, the sensitivity and specificity should equal 1