In silico predictions of target clinical efficacy

As technological advances revolutionize the process of novel target identification in drug discovery, the problem of validating this ever-growing number of targets against predicted clinical efficacy in humans is creating a bottleneck. All methods of novel target identification rely on partial and isolated models of human disease. For example, methods such as differential gene expression (comparing the upregulation of a particular gene in several sick versus healthy patients) and high-throughput compound screening (identifying compounds that hit a pathway that is thought to be involved in a disease process) are important research that intimate target involvement in a particular disease process, but such ‘hints’ lack specificity for predicting the clinical efficacy of a target. Given that current target identification methods are an imperfect predictor of clinical efficacy and that moving all targets forward through to development is prohibitive in terms of cost and time - how can rational choices between novel targets be made?