Saving Lives: Constructing a Nationwide Cohort With Multi-modal Data to Improve Precision in Prediction and Prevention of Suicide

Suicide is a major public health issue, causing severe impact on individuals and families, as well as relevant societal costs. In Sweden alone, ~70,000 years of potential life are lost each year due to suicide. Despite dedicated research efforts and prevention strategies, suicide-related outcomes are still difficult to predict and prevent. The goal of this research program is to utilize unique resources in Sweden to improve prediction of suicide by integrating environmental factors captured by national registers and genetic information using multi-modal modelling. In aim 1, we will create a large suicide biobank by collecting neonatal blood spots stored at the Swedish PKU biobank from all individuals who died by suicide, i.e., total coverage of all suicides in Sweden from individuals born 1975 and onwards (n=5000). DNA will be extracted from each of these blood spots. Data from the national registers will be added, covering major risk factors across the lifetime (socioeconomic, demographic, and medical). Further, blood spots and register data from 10,000 matched controls with no suicidal outcomes will be available through collaboration. In aim 2, we will identify register-based risk factors and genetic variants associated with suicide by genotyping all DNA samples from cases and matched controls and performing a case-control genome-wide association study meta-analysis. In aim 3, we will combine hundreds of candidate suicide predictors from the national registers (covering demographics, socioeconomic status, electronic medical records, criminality) with genetic predictors (polygenic risk scores for suicide, depression, impulsivity, and substance misuse) using both established quantitative modelling and newer machine learning approaches. This study should help provide a refined understanding of factors influencing suicide and an increased ability to identify those at risk.