2019-11-19

023_ Genetic and phenotypic determinants of disease onset in bipolar disorder.

Research Question and Aims

The diagnosis of Bipolar Disorder (BD) can be extremely challenging in clinical practice, among others due to an unspecific prodromal state and frequent misdiagnosis with major depressive disorder (depression-first BD). Gaining knowledge on the genetic and phenotypic factors associated with an early age at onset (AAO) and a specific polarity at onset (PAO) could inform the development of future diagnostic tools and eventually accelerate diagnosis.
Early disease onset is associated with an unfavorable clinical outcome and constitutes a clinically and biologically homogenous subgroup within BD. Previous studies have found an accumulation of early AAO in BD families and have therefore hypothesized a larger genetic contribution to the early onset cases. However, GWASs on AAO (Belmonte Mahon et al., 2011; Jamain et al., 2014; Nassan et al., 2015) found no GWS loci. Moreover, our recent analysis investigating the association between BD- and SCZ-PRS and AAO in a sample of 1,995 BD-1 patients also reported negative results. (Kalman et al., 2018) However, Ruderfer et al 2018 using a significantly larger sample, found a strong trend for negative correlation between AAO and SCZ-PRS. (Ruderfer et al., 2018) Thanks to coordinated international efforts (e.g. PGC, ConLiGen) the number of BD patients with available genotype and phenotype information is constantly growing. Thus, there is momentum now to conduct larger GWAS analyses and investigate the genetic architecture of AAO, PAO.

Analytic Plan

We have already successfully applied to the PGC-BD, ConLiGen and FOR2107 and gained access to the AAO and PAO information of BD. We aim to perform (1) four GWAS studies: (1.1.) investigating AAO as continuous phenotype in both a cross-diagnostic and (1.2.) diagnosis-specific setting (1.3.) investigating AAO as a categorical phenotype and (1.4) comparing depression first BD and mania first BD. Furthermore, (2) we aim to investigate the functional and biological annotation of the SNPs and conduct gene-set analyses using FUMA GWAS and MAGMA, respectively. (3) Additionally, we aim to investigate the association of publicly available PRSs (e.g. SCZ-PRS) with the phenotypes of interest.
We will perform QC as part of the standard Ricopili protocol. Plink1.9 will be used for the GWAS analyses and for calculating PRS. Downstream analyses will be carried out using R or SPSS. The potential effects of population stratification will be addressed by using principal components obtained with EIGENSTRAT and including them in the model.
In the phenotypic analyses we aim to investigate (using elastic net regression) whether certain "pre-AAO" characteristics of the patients (e.g. childhood trauma, migration status - see full variable list) predict a specific type of onset. Furthermore, we wish to investigate whether a specific illness onset (e.g. early-AAO, PAO-M, PAO-D) is associated with a distinct disease course (i.e. presence of psychotic symptoms, number of hospitalizations - see full variable list).

Resources needed

Phenotypes:
"v1_id", "v1_center", "v1_sex", "v1_ageBL", "v1_marital_stat", "v1_partner", "v1_liv_aln","v1_ed_status", "v1_curr_paid_empl", "v1_disabl_pens", "v1_spec_emp", "v1_wrk_abs_pst_5_yrs","v1_dur_illness", "v1_tms_daypat_outpat_trm", "v1_cat_daypat_outpat_trm", "v1_scid_dsm_dx_cat", "v1_scid_no_MDE", "v1_scid_no_mania", "v1_scid_no_hypomania", "v4_opcrit", "gwas_id"; v1_scid_evr_suic_ide; v1_scid_suic_ide; v1_suic_attmpt; v1_scid_no_suic_attmpt; v1_ever_delus; v1_ever_halls; v1_scid_ever_psyc; v1_ever_smkd; v1_alc_pst12_mths; v1_alc_5orm; v1_lftm_alc_dep; v1_evr_ill_drg; v4_alda_*; v3_cts_*; v1_nrpsy_mwtb; v1_fam_hist; v1_big_five_*; v1_cntr_brth*; v1_age_1st_out_trm; v1_age_1st_inpat_trm; v1_scid_age_MDE; v1_scid_age_mania; v1_clin_scid_age_hypomania

Genotypes:
First to calculate GRS and ancestry components; on the long-term for GWAS
Analysis will be conducted on local computers and at the Leibniz supercomputing centre (LRZ).
The validation of findings and the long-term aims require the inclusion of further cohorts, which have, so far, not been contacted regarding the present proposal.