Suicidal behavior is a complex phenomenon, ranging from low-lethality, low-intent impulsive acts to high- lethality high-intent suicidal acts, and thus likely to be associated with multiple underlying subtypes. Genetic associates of suicidal behavior have been identified in several studies, yet the effect sizes for are usually modest, possibly because of the heterogeneity in the suicidal population and the behavior. Using the Million Veteran Program Gamma computational platform, we propose a unique combination of statistical and machine learning methods to develop our subtypes based on Electronic Patient Records and self-reports; followed by a careful genomic analysis of the resulting subtypes compared to two control groups chosen from the same cohort, with and without mental health disorders. This project aims to develop sophisticated diagnosis tools for preventing future suicidal behavior in US Veterans at high risk. Moreover, the biomarkers identified from this study will be directly applied for validation in the PI and co-investigators? longitudinal study of high-risk VA patients, a natural validation sample from the same population; promising a combination of the power of both studies.
Suicide is a major health crisis within our Veteran population. With the Million Veteran Project, we plan to develop computational tools to identify veterans at high risk of suicide. Using sophisticated data mining methods integrating both clinical and genetic data, the goal is to identify bio-signatures of suicidality that can be used for monitoring and prevention of Veteran suicides.