Polypharmacology, which focuses on designing drugs to target multiple receptors, has emerged as a new paradigm in drug discovery. To rationally design multi-target drugs, it is fundamental to understand protein-ligand interactions on a proteome scale. We have developed a Proteome-wide Off-target Pipeline (POP) that integrates ligand binding site analysis, protein-ligand docking, the statistical analysis of docking scores, and electrostatic potential calculations. The utility of POP is demonstrated by a case study, in which the molecular mechanism of anti-cancer effect of Nelfinavir is hypothesized. By combining structural proteome-wide off-target identification and systems biology, it is possible for us to correlate drug perturbations with clinical outcomes.