CausalPath (
www.causalpath.org) analysis of 2‐h drug/vehicle treatment data. Log
2 (L/H PAR) for all 6 WHIM models was used for this analysis. Moderated one‐sample
t‐test was used to analyze 2‐h treatment data. Resulting table was used for the CausalPath analysis. CausalPath network generated by comparing drug‐treated PDX samples to the controls at 2 h. Nodes represent proteins, and the (p) labels on the nodes represent significant differences in site‐specific phosphopeptide measurements. (p) Blue background color indicates a downregulated site, red background color indicates an upregulated site. Green border color around (p) indicates an activatory site, and a red border color indicates an inhibitory site. Green edges represent known site‐specific phosphorylations, and red edges represent dephosphorylations. The label (i) indicates an inhibited protein. In the case of PI3KCA, the label (i) indicates our manually inserted hypothesis of inactivated PIK3CA due to the drug effect. All other (i) labels on the graph are generated automatically by the CausalPath algorithm through statistical evaluation of the changes at the downstream of the protein. CausalPath infers the PI3KCA ‐> AKT1 relation, indicating the downregulated phosphorylation of AKT1 is likely due to inhibition of PIK3CA. Additionally, statistical measurements on the downstream of AKT proteins indicate their inactivation. We observe that this effect extends over downstream targets of AKT such as mTOR.