Background: Motility-related gastrointestinal (GI) adverse drug reactions (GADRs) such as diarrhea and constipation are a common and deleterious feature associated with drug development. Novel biomarkers of GI function are therefore required to aid decision making on the GI liability of compounds in development. Methods: Fifteen compounds associated with or without clinical GADRs were used to assess the ability of an in vitro colonic motility bioassay to predict motility-related GADRs. Compounds were examined in a blinded fashion for their effects on mouse colonic peristaltic motor complexes in vitro. For each compound concentration-response relationships were determined and the results compared to clinical data. Compounds were also assessed using GI transit measurements obtained using an in vivo rat charcoal meal model. Key Results: Within a clinically relevant dosing range, the in vitro assay identified five true and three false positives, four true and three false negatives, which gave a predictive capacity of 60%. The in vivo assay detected four true and four false positives, four false and three true negatives, giving rise to a predictive capacity for this model of 47%. Conclusions & Inferences: Overall these results imply that both assays are poor predictors of GADRs. Further analysis would benefit from a larger compound set, but the data show a clear need for improved models for use in safety pharmacology assessment of GI motility.