Discovering Venetoclax Resistance Predictors

First, each attribute was associated to standardized (min-max normalized) Venetoclax resistance from BEAT-AML study at univariate settings using lm function in R environment (v4.1.0) (Supplementary Table S12). Then, p-values and coefficients were calculated using summary function and then p-values were multiple hypothesis corrected using Benjamini-Hochberg procedure. For multivariate models, FAB classification, and ECC levels (except CD14+ Monocytes) were excluded as only 76 out of 460 samples of BEAT-AML had FAB classifications and as other ECCs are not independent of CD14+ Monocyte percentages (as the question is whether CD14+ Monocyte levels are independently predictive of Venetoclax resistance given BCL-2 expression in the same model). Again, p-values and coefficients were calculated with summary function and plotted in a volcano plot (Fig. 3f, Supplementary Table S13).