Project Title: Identifying suicidality subtypes using machine learning and genomic data
Total Awarded: $200,000
Abstract: Background: There are over 800,000 suicide deaths worldwide each year. With over 3,500 suicides and 70,000 suicide attempts each year, it is the ninth leading cause of mortality in Canada. Suicidal thinking (ideation) and behaviour (SIB) is very complex and is likely affected by many variables, including genetic factors. Finding genetic factors for SIB may require separating SIB into subtypes to reduce complexity. Objectives: We aim to collect data from 550 participants and evaluate them on their SIB and a number of endophenotypes. We aim to uncover SIB subgroups based on these endophenotype measures. Then, we aim to find genes relating to each of the SIB subgroups.Hypotheses: (1) Patients cluster into SIB subtypes based on the endophenotypes. (2) Genetic variants, in the form of genes, gene-sets, and risk scores calculated from thousands of genes, each with small effect, are associated with SIB subtypes. Methods: Participants will be assessed through a series of questionnaires about their history of SIB (Columbia Suicide Severity Rating Scale (CSSRS)), stressful life events, personality, as well as through several neurocognitive tasks. A number of machine learning algorithms will be used to cluster participants based on these clinical assessments. A preliminary genetic analysis (gene-based, gene-set, and multigene-risk score) will be conducted across these SIB subtypes.
Significance: Findings from this study will provide for a better understanding of the biology of one or more SIB subtypes, which will inform the design of future suicide studies and the development of subtype-specific drug therapies and prevention.