In spite of major efforts and investment in development of psychiatric drugs, many clinical trials have failed in recent decades, and clinicians still prescribe drugs that were discovered many years ago. Although multiple reasons have been discussed for the drug development deadlock, we focus here on one of the major possible biological reasons: differences between the characteristics of drug targets in preclinical models and the corresponding targets in patients. Importantly, based on technological advances in single-cell analysis, we propose here a framework for the use of available and newly emerging knowledge from single-cell and spatial omics studies to evaluate and potentially improve the translational predictivity of preclinical models before commencing preclinical and, in particular, clinical studies. We believe that these recommendations will improve preclinical models and the ability to assess drugs in clinical trials, reducing failure rates in expensive late-stage trials and ultimately benefitting psychiatric drug discovery and development.