The genomic stratification of schizophrenia

Common genetic variation is integral to the genetic architecture of schizophrenia, but until recently, the sample sizes and computational advances needed to detect them did not exist. Through the formation of international consortia such as the PGC, variants associated with schizophrenia can now be robustly detected, however it has come at the cost of phenotypic detail. As a result, genetics specific to more homogenous patient groups, for example individuals with treatment resistance, have remained a challenge to identify. In addition, the heterogenous nature of schizophrenia cohorts, the disorder itself, and the significant overlap with patients of other major psychiatric disorders, has made the identification of 'schizophrenia-unique' variation and neurobiology challenging. To achieve more personalised medicine approaches in schizophrenia, the stratification of individuals into more genetically and phenotypically homogenous groups will be vital. I first examined the genetic differences between schizophrenia and bipolar disorder, using a recently published research method, the CC-GWAS. I identified 27 loci that were differentially associated with schizophrenia and bipolar disorder, with follow up interrogation of the summary statistics pointing to 26 of them being 'schizophrenia-unique' loci. Following this I shifted the focus from cross-disorder to within, in an attempt to identify common genetic variation specific to treatment-resistant schizophrenia (TRS). I performed a direct case-case GWAS of ~40,000 individuals with TRS and non-TRS, identifying a genomewide significant locus that was positively associated with TRS. iii Finally, I investigated the relationship between clozapine and neutrophil counts, identifying a significant association between clozapine:norclozapine ratio and absolute neutrophil count, as well as associations with a small set of pharmacogenomic variants associated with the metabolism of clozapine.