Exploring the implications of case selection methods for psychiatric molecular genetic studies

Researchers selecting probands for molecular genetic studies confront a range of sampling issues with modest empirical guidance. In this paper, using cases of major depression (MD), anxiety disorders (AD) alcohol use disorder (AUD), drug use disorder (DUD), bipolar disorder (BD) and schizophrenia (SZ) from a large population cohort of all native Swedes born 1940-2003, we examine the implications of three proband selection decisions by exploring profiles of genetic risks assessed using the validated family genetic risk scores. The impact of censoring cases with comorbid diagnoses is quite variable, depending on the frequency of that disorder in the case sample and the genetic relationship of the censored to the primary disorder. In an MD cohort, censoring SZ cases produces only a focal small decrease in schizophrenia genetic risk while censoring AD cases produces a wide-spread reduction in genetic risk for MD and most other disorders. We examine the value of censoring cases of SZ, BD and MD whose onset was preceded by one to two years by first episodes of DUD or AUD. We do not see any increase in genetic risk for these "screened" cohorts. Secondary ascertainment, where disorder A is ascertained as a comorbid diagnosis in a sample collected for disorder B, can, in certain situations, produces large increases in the genetic risk for disorder B and associated disorders in cases of A. However, if disorder B is closely genetically related to disorder A (as seen with MD/AD and DUD/AUD pairings), the pattern differs dramatically and produces a general moderate elevation across the genetic risk profile. These findings provide guidelines for future investigators and suggest caution when screening out comorbid disorders and when utilizing secondary ascertainment.