The question of whether toxicity caused by small molecules is due to on-target or off-target pharmacological effects is frequently encountered in drug discovery programs and of decision making significance. Genetically engineered mouse models are widely used to assess the safety consequences of depleting a target, but have limitations in predicting the toxic effects of target-related small molecule drug action. Diverse chemical analogs were designed/selected to investigate the root cause of cardiovascular toxicity encountered in a program aimed at discovering inhibitors of pan-Group I p21-activated kinases (PAK1, 2, and 3) for use in breast cancer. Mouse tolerability studies with these compounds revealed persistent toxicity, a correlation of minimum toxic concentrations and PAK1/2 mediated cellular potencies, and absence of toxicity with structural analogs devoid of PAK1/2 activity. Our data suggest that the toxicity results from the inhibition of PAK2, which may be enhanced by PAK1 inhibition, and caution against continued pursuit of pan-Group I PAK inhibitors in drug discovery. Carefully selected sets of small molecule tool compounds have significant value in probing the safety of drug targets, and their use in tolerability experiments is highly complementary to gene knockout studies.