Fig. 1.

High-level overview of this work. (A) The goal of this work is to infer a topology of relationships between pharmacogenomics datasets (an example topology is illustrated here). (B) When two datasets share information (i.e. when their RV coefficient is non-zero), we will indicate them as connected in a topology. (C) A topology of three datasets that all share information. We will convert this topology to the one depicted in (D) if the shared information between X₁ and X₃ is fully contained in X₂. (E) To create these topologies we will draw on methods for inferring a topology between single variables using partial correlations. Top: the original causality graph. Middle: the topology as inferred using correlations. Bottom: the inferred topology using partial correlations