Fig. 6

Gradient boosting machine models accurately discriminate CTRL vs. AD astrocytes and microglia. Receiver operating characteristic (ROC) curves demonstrate the high discriminative power of the gradient boosting machine (GBM) models to discern between CTRL and AD (a) astrocytes and (b) microglia based on mean gray intensity (MGI) data from thousands of high-plex single-cell profiles. Compare with the performance of GBM models trained on single marker intensity data, namely (a) GFAP and (b) MHC2 or CD68. Rankings of the variable importance scores shown in the horizontal bar plots reveal the most relevant markers for each classification task, respectively