Over the past 20 years a large database has been developed that shows a number of biological abnormalities that distinguish depressed patients from healthy volunteers. Unfortunately, the diagnostic criteria do not include such changes since they do not seem to be of clinical relevance. One of the reasons frequently mentioned in the literature is that these biological abnormalities are not specific to the different diagnostic entities. One of the reasons that biomedical tests are not able to diagnose DSM or ICD defined disorders is because specific biological measurements are evident across diagnostic categories. Therefore, it seems to be logical that research should be focused on biological homogenous subtypes across phenotypic diagnoses. This notion is also backed up by practical pharmacology in the sense that medications with antidepressant properties are used in depression, anxiety disorders, schizophrenia as well as in dementia. On the other hand the group of atypical antipsychotics have been demonstrated to show efficacy not only in schizophrenia but also in bipolar disorder, partly in anxiety disorder as well as in specific symptoms of dementia. To validate specific biomarker-defined subtypes regardless of the DSM or ICD diagnoses will help to understand the underlying biology in the acute and very likely also long term course of the different new to define illnesses. This approach could speak for a stratified psychiatric approach that is the scientific expression of the often erroneously used term of personalised medicine which could be interpreted that there is only one medicine for one person which seems to be highly unlikely.