Various classes of antidepressants have been used in the treatment of major depressive disorder (MDD); however, treatment efficacy is inadequate, as 30-40% of patients do not experience response even after sufficiently long treatment period with adequate dose of antidepressant. For the treatment-resistant patient to the therapy based on the generalized evidence, is it possible to provide an appropriate and improved treatment based on personalized medicine, taking into account predictable candidates such as sub-symptoms of depression and genetic factors instead? There is only little evidence for this in Japanese MDD, and consequently we use the evidence of Caucasians as reference, however, could we use the evidence of the population whose genetical, social, and cultural background are very different from Japanese population?
 In this review, I will refer to our randomized controlled studies that have some predictable candidates including genetic factors designed for personalized medicine in MDD patients, and present an overview of procedures for making predictions of current treatment and proceeding towards personalized medicine.
 <Author's abstract>