Pathomechanisms and Signatures in the Longitudinal Course of Psychosis



029_ Analysis of the influence of polygenic risk on lifetime cannabis use in PsyCourse cohort

Research Question and Aims

It is well established that cannabis abuse is much more prevalent among patients with psychiatric disorders than in the general population. Genetic studies have also revealed that a genetic (polygenic) predisposition to schizophrenia might be associated with increased use of cannabis in healthy individuals (Power 2014; Verweij 2017; Reginsson 2018). However less is know about the influence of polygenic risk on cannabis use in populations of psychiatric patients. Our unpublished preliminary data in an independent dataset has already shown that genetic risk for schizophrenia plays a role in the lifetime risk of cannabis use in bipolar disorder patients.

In the light of the evidence from these studies, here we propose to generate polygenic scores regarding schizophrenia, bipolar disorder, lifetime cannabis use, and cannabis use disorder within the PsyCourse cohort and estimate their association with cannabis-use related variables, both transdiagnostically and in a diagnostic-specific fashion.

Analytic Plan

General hypothesis:
Polygenic risk of schizophrenia, bipolar disorder, lifetime cannabis use, and/or cannabis use disorder contribute to the risk of cannabis use in psychiatric patients and healthy controls.

Data from all samples of patients and healthy controls in PsyCourse who have genotype/cannabis use phenotype data available will be included in this study.

Statistical analysis:
PLINK 1.90 will be used for polygenic score calculation. The most recent GWASes of schizophrenia, bipolar disorder, lifetime cannabis use, and cannabis use disorder will constitute the discovery sample. Polygenic scores will be calculated based on summary statistics from the discovery dataset excluding rare SNPs (MAF < 0.5%), low quality imputed variants (info score <90%), indels, ambiguous markers (A/T and C/G). Data will be clumped in windows of 1000 kbp, discarding variants in LD (R2>.2). Scores will be calculated based on p value thresholds ranging from p < 5 x 10-8 to p < 1. In addition, continuous shrinkage PRS (PRScs) will also be calculated given their improved performance with respect to P-value thresholding + clumping methodologies. Imputed genotypes will be used for these calculations.

Statistical analysis:
The association of polygenic scores with relevant cannabis use variables will be studied using linear/logistic models adjusting for sex, age, treatment, ancestry components, and recruitment site. The analyses will be carried independently for each major diagnosis available in PsyCourse (schizophrenia, bipolar disorder, major depressive disorder, and also healthy controls. Subsequently, an analysis including all subjects regardless of their status will also be performed.

Resources needed

Diagnosis/status (v1_scid_dsm_dx_cat)
Age at first interview (continuous [years], v1_age)
Sex (dichotomous [M,F], v1_sex)
Recruitment center variable (categorical [see below], v1_center)

Raw data on illicit drugs:

Raw genotypes pre-imputation to calculate PCAs
Imputed genotypes