Genome-wide association studies (GWAS) have recently been very successful in identifying common genomic variants as risk factors for some complex diseases, such as diabetes mellitus and macular degeneration. However, the identification of genomic abnormalities in psychiatric disorders remains challenging. Our proposal is aimed at addressing some of the issues that might hinder the identification of genetic risk factors for psychiatric disorders. In a combined analysis of publicly available large data sets of the NIMH, we will test the HYPOTHESIS that assigning individuals to specific subgroups based on co-morbid conditions will reveal groups of patients that are more homogeneous with regard to their genetic risk factors than the too wide classification of simply """"""""bipolar disorder"""""""".
Specific aims : 1. Latent class clustering will be used to explore the so far unobserved heterogeneity in samples of bipolar patients 2. Class membership probabilities for the subgroups will then be used as phenotype in case/control association studies, as well as in copy-number association studies, utilizing existing genome-wide SNP data. We will determine if the subgroups show stronger association with genomic variants than the diagnostic category """"""""bipolar disorder"""""""". 3. Throughout the project, we will genotype additional samples from the NIMH genetic initiatives in order to increase the sample size for our subclasses. These additional samples will be incorporated into our case/control association analyses. 4. We will then examine, if SNPs and genomic variants that are significantly associated with latent class membership probabilities are located in or near genes, and if those genes aggregate in specific signaling and metabolic pathways using bioinformatics tools and databases. 5. Our analysis will first focus on bipolar disorder, and then we will also explore model-based clustering and association with class membership probabilities in schizophrenia in order to determine possible common genetic risk factors. Significance: Our study will explore the possibility of unobserved heterogeneity in psychiatric patients based on common co-morbid conditions. We propose that taking heterogeneity into account will improve the results of genome-wide association studies in psychiatric disorders. Identifying genomic variants associated with certain clinical subgroups of patients would allow a better understanding of the underlying patho-physiology of these disorders, and may facilitate early diagnosis and interventions.
Our study will address the variability in psychiatric disorders. We will focus on common co-morbid conditions in bipolar disorder and schizophrenia and their relationship to genetic risk variants. Identifying genomic variants associated with certain clinical subgroups of patients would improve our understanding of the underlying patho-physiology of these disorders;it would also help in early diagnosis and more specific treatment.