Leveraging genome-wide association and clinical data in revealing schizophrenia subgroups

Publication date: Available online 22 September 2018Source: Journal of Psychiatric ResearchAuthor(s): Liangying Yin, Eric Fuk-Chi Cheung, Ronald Yuk-Lun Chen, Emily Hoi-Man Wong, Pak-Chung Sham, Hon-Cheong SoAbstractSchizophrenia(SCZ) has long been recognized as a highly heterogeneous disorder. Patients differed in their clinical manifestations, prognosis, and underlying pathophysiologies. Here we presented and applied a framework for finding subtypes of SCZ utilizing genome-wide association study(GWAS) and clinical data. We postulated that genetic information may help stratify patient into useful subgroups, and incorporation of other clinical information and cognitive profiles will further improve patient subtyping. We conducted cluster analysis in 387 Hong Kong Chinese with SCZ. First we performed ‘single-view’ clustering using genetic or clinical data alone, then proceeded to ‘multi-view’ clustering (MVC) accounting for both types of information. We validated clustering results by assessing subgroup differences in various outcomes. We found significant differences in outcomes including treatment response, disease course and symptom severity (Simes overall p-value using MVC = 1.64E-9). Overall speaking, we identified three subgroups with good, intermediate and poor prognosis respectively. MVC generally out-performed single-view methods. The analysis was repeated for different sets of input SNPs, and stratified analysis of male and female patients, and the result...
Source: Journal of Psychiatric Research - Category: Psychiatry Source Type: research