Identifying biological and clinical markers of treatment response in depression is an area of intense research that holds promise for increasing the efficiency and efficacy of resolving a major depressive episode and preventing future episodes. Collateral benefits include decreased healthcare costs and increased workplace productivity. Despite research advances in many areas, efforts to identify biomarkers have not revealed any consistently validated candidates. Studies of clinical characteristics, genetic, neuroimaging, and various biochemical markers have all shown promise in discrete studies, but these findings have not translated into a personalized medicine approach to treating individual patients in the clinic. We propose that an integrated study of a range of biomarker candidates from across different modalities is required. Furthermore, advanced mathematical modeling and pattern recognition methods are required to detect important biological signatures associated with treatment outcome. Through an informatics-based integration of the various clinical, molecular and imaging parameters that are known to be important in the pathophysiology of depression, it becomes possible to encompass the complexity of contributing factors and phenotypic presentations of depression, and identify the key signatures of treatment response.