Major depressive disorder (MDD) is a serious problem in public health and has been the subject of a massive to identify gene-environment interactions (GxE) (i.e., when stressors "trigger" pre-existing genetic vulnerabilities to increase risk), with a goal of shedding light on new inroads for treatment or prevention of MDD. However, progress in the GxE field has been stymied by extremely inconsistent findings, as highlighted by a recent meta-analysis (Risch et al., 2009). This proposal examines several possible causes for these inconsistent findings. One issue, as highlighted by a landmark critique of this research area (Monroe & Reid, 2008), is that GxE reports on MDD have operationalized life stress in highly inconsistent ways (e.g., focusing on the total number of stressors experienced or the severity rating of the worst event)-to the extent that no two reports examine life stress in precisely the same manner. This may have occurred because no one has yet examined in a systematic fashion what operationalization of life stress best predicts MDD onset. Further, whether chronic stress or childhood adversity should be included in GxE tests is not known, as almost no research directly compares their predictive power with that of the traditionally studied environmental pathogen-stressful life events. A second issue is that no GxE reports on MDD examine whether the MDD cases under study are first lifetime onsets of MDD or recurrences. This is important because it is well- established that the relationship of severe stressors to MDD episode onset degrades with successive recurrences; one hypothesis is that MDD sensitizes individuals to the effects of stress, and later episodes can be "triggered" by progressively milder stressors. The larger the proportion of recurrences of MDD in a report, the lower the probability of finding a genuine GxE interaction would be. Third and finally, gender and race/ethnicity may moderate the relationship between life stress and MDD onset. Accounting for such a moderator may further reduce statistical error variance, improving the probability of findings GxE interactions. These issues will be examined and the results applied in GxE testing (in previously studied and novel polymorphisms) using existing data from a moderately large, racially diverse, prospective longitudinal investigation of risk for emotional disorders in late adolescence to early adulthood. Proportional hazards regression (survival) analysis and multilevel modeling will be employed. In keeping with the mission of the sponsoring agency, such a plan of research may: 1) remove barriers to further identifying the genetic and environmental factors associated with MDD, 2) inform the relative importance of different types of life stress for increasing risk for MDD, and 3) enhance knowledge of how gender and race/ethnicity may influence the onset and recurrence of MDD through differences in sensitivity to life stress.