The psychiatric genomics consortium sub-phenotype project

The PGC bipolar disorder working group (PGC-BD) has long recognized the importance of research into clinical sub-phenotypes in the GWAS efforts of bipolar disorder. Previous studies of BD sub-phenotypes have contributed to a better understanding of the biology of bipolar disorder and its relationship with other psychiatric disorders such as schizophrenia. In the last two years, the efforts have been refocused on additional sub-phenotypes and revitalized the data collection from more cohorts contributing to the PGC-BD. Here, we aim to present the preliminary results of the BD sub-phenotypes latest data. Sub-phenotypes were identified and prioritized through discussions and a consensus was reached by PGC-BD. Sub-phenotype definitions and harmonization efforts were performed by a Clinical Expert Group. Sub-phenotype Data Management Team set up data storage and a computing workspace committed to sub-phenotype data was set up to ensure data safety. After the data collection template was prepared, requests for sub-phenotype data were made to all investigators who contributed samples for BD GWAS. The collected data were cleaned, coded, and analyzed using R. Correlations and t-tests were performed as applicable. Genetic correlations were investigated by employing the Genome-wide Complex Trait Analyses (GCTA) framework for selected sub-phenotypes (n= 4,311 - 8,852). Data were obtained from 40 of the 57 cohorts constituting 21,133 individuals of European Ancestry with bipolar disorder. Cohorts widely varied in sample size and availability of specific sub-phenotype data. There was a significant increase in data availability: BD subtype (18,764), age of onset (17,647), history of psychosis (15,244), alcohol and drug use (12,545), suicide attempt (12,060), and panic disorder (10,567). Among others, earlier age of onset of BD was associated with alcohol or drug use, suicide attempts, and comorbid panic disorder (P < 2.2 × 10E-16). Furthermore, the age of onset had negative genetic correlations with psychosis (-0.13 ± 0.05) and rapid cycling (-0.24 ± 0.04). Positive genetic correlations were found between suicidal ideation and panic disorder (0.22 ± 0.08), and suicide attempts and rapid cycling (0.14 ± 0.04). Future efforts will continue to improve the quality and power of sub-phenotype data by strengthening collaborative efforts. Additional efforts to include populations with diverse ancestry, strategies to obtain data from biobanks, and improved harmonization of sub-phenotypes will enhance discovery and ultimately the clinical care of individuals with bipolar disorder.