Mechanistic models of signal transduction have emerged as valuable tools for untangling complex signaling networks and gaining detailed insight into pathway dynamics. The natural extension of these tools is for the design of therapeutic strategies. We have generated a novel computational model of lipopolysaccharide-induced p38 signaling in the context of TNF-alpha production in inflammatory disease. Using experimental measurement of protein levels and phospho-protein time courses, populations of model parameters were estimated. With a collection of parameter sets, reflecting virtual diversity, we step through analysis of the p38 signaling pathway model to answer specific drug discovery questions regarding target prioritization, inhibitor simulation, model robustness and co-drugging. We demonstrate that target selection cannot be assessed independently from inhibitor mechanism of action and is also linked with robustness to cellular variability. Finally, we assert that in the face of parameter uncertainty one can still uncover consistent findings that can guide drug discovery efforts.