Large population based studies demonstrate a link between childhood trauma (CT) and increased prevalence of a wide range of psychiatric disorders, including mood disorders and psychosis. The identification of CT as an environmental risk factor for bipolar disorders (BD) and schizophrenia is of crucial importance to move to GxE interaction studies. In this presentation, we will mainly use the example of bipolar disorders (BD) and draw some parallels with psychosis. First we have demonstrated that CT were associated to BD using case-control studies (with a dose-effect of emotional abuse), as this has been previously demonstrated in psychosis by meta-analytic approaches. We also have shown that CT (mainly emotional and sexual abuses) influenced the clinical expression and course of BD, with associations between CT, earlier age at onset, rapid cycling and suicidal behavior. Some of these associations were also observed in schizophrenia and thus appeared as relatively non-specific. Moreover some cumulative effects of CT and cannabis misuse on the age at onset of BD or on the transition to psychosis have been suggested. Using psychopathological dimensions as outputs, we have been able to demonstrate that CT influence affective regulation and impulsivity-related dimensions, that could mediate the links between CT and clinical categorical outputs. Disentangling these pathways from CT, through dimensions, to clinical expression could shed some light on the clinical and dimensional aspects to could be incorporated in future GenexCT interaction studies. As an example, we will present preliminary data on the interaction between CT and serotonin transporter gene variants but also dysimmunity-related genes variants on the age at onset in BD.