Fig. 4: Propagating attributions to gene sets enables higher-level understanding.
From: Explaining a series of models by propagating Shapley values
a Gene and gene set attributions for predicting Alzheimer's disease using gene expression data. b Gene and gene set attributions for predicting breast cancer tumor stage using gene expression data. Residuals in the gene set attributions summarize contributions for genes that are not present in any gene set and describes variations in output not described by the pathways we analyzed. Note that (a) and (b) show summary plots (Supplementary Methods Section 1.3.3).
