Achieving effective gas exchange in the mammalian lung requires coordinated matching of ventilation and perfusion, both of which are sensitive to the branching structure of the lung’s airway and vascular trees, hydrostatic pressure, and gravity acting to deform the lung within the chest wall. Virtually all pathology of the lung impairs gas exchange by disrupting ventilation-perfusion (V/Q) matching, but because of the complexity of the integrated respiratory system it is not always possible to predict how impairment to a single mechanism will affect the function of the system as a whole. To understand the emergence of asthmatic bronchoconstriction and variable patient response to acute pulmonary embolism (APE) we have developed computational methods that exploit volumetric computed tomography – from normal volunteers, clinical subjects with suspected APE, and animal validation studies – and pulmonary function testing, to derive structure-based models for predicting function in the individual. These data-driven multi-scale models include the pathways for ventilation and perfusion and their interaction at the gas exchange surface, and calcium signaling for smooth muscle force development. Multi-scale analysis reveals synergistic interactions between deep inspirations and smooth muscle fluidization during bronchoconstriction, which may contribute to different responses in normals and asthmatics; and that the cardio-respiratory response to mitigate hypoxia in APE is dependent on the location of embolus, with our model predicting V/Q disruption in regions that are distant from the embolus, and the severity of predicted response tightly correlated with clinically identified right ventricular dysfunction.