Fig.1

Workflow of analyses and features in MAVEN which requires 3 main inputs: Targets (A, optional) which can be known/hypothesised or predicted from the compound structure with PIDGINv4, [4, 19] a signed and directed prior knowledge network for causal reasoning B and differential gene expression signatures as e.g., log2FC, t-statistic C which is used for the inference of transcription factor and pathway activities with DoRothEA [20] and PROGENy [21]. These are used as input to CARNIVAL [7] for optimisation of the signalling network, connecting the targets to the TFs via intermediate signalling proteins D. The resulting predicted signalling network is displayed in the GUI and interactive pathway enrichment can be carried out to contextualise the signalling proteins and interpret the compound’s mechanism of action E. Figure adapted with permission from https://github.com/saezlab/carnival