A CHROMATIN CATALOG FOR THE INTERPRETATION OF GENETIC ASSOCIATIONS OF PSYCHIATRIC DISORDERS

Background Schizophrenia is an often devastating psychiatric disorder with substantial morbidity, mortality, and personal and societal costs. Genome-wide association, Copy Number Variant (CNV), and exome studies have demonstrated that schizophrenia is a complex genetic disorder, with likely hundreds of genes contributing to its genetic architecture. For most schizophrenia GWAS loci, we are unable to pinpoint the specific genes that connect with these GWAS signals or the direction of association. This markedly limits the biological, clinical, and therapeutic utility of these findings. Prior studies have demonstrated the importance of chromatin looping as a regulatory mechanism and as a way to connect Single Nucleotide Polymorphisms (SNPs) to genes. We aim to facilitate the interpretation and prioritization of genomic findings from schizophrenia and other psychiatric disorders by generating a catalog of chromatin interactions in fetal and adult brain using easy Hi-C (eHi-C, a variation of Hi-C). Generating CNS-specific functional genomics data is critical for the development of a mechanistic understanding of genetic findings in psychiatric disorders. Methods Chromosome conformation capture methods enable the identification of chromatin interactions in vivo. In this study, we implemented eHi-C to generate a comprehensive map of brain-specific chromatin interactions in fetal (n=3) and adult frontal cortex (n=3) at a genome-wide level (~1 billion reads/sample). We used in-house pipelines to process the Hi-C data in line with established methods. As part of quality control, we examined summary statistics of Hi-C reads, including total number of reads, total number of uniquely mapped reads, total number of intra-chromosomal reads, and total number of intra-chromosomal reads which are >15 kb. HiCNorm was used to normalize raw Hi-C contact matrices. We used the insulation square method to identify the boundary regions of topologically associating domains. We used Fit-Hi-C to detect chromatin interaction peaks for all intra-chromosomal interactions within 2 Mb. We further applied HiCNormCis to identify the frequently interacting regions (FIREs), and performed GREAT analysis to assign biological meanings to the detected FIREs. We used HUGIn to visualize Hi-C data and related genetic and epigenetic features. Results All summary statistics from both fetal and adult samples were comparable to previous Hi-C studies. Biological replicates for each sample, showed high reproducibility (Pearson correlation coefficients > 0.96). We were able to connect several loci to genes for the recent PGC Major Depressive Disorder (MDD) GWAS that identified 44 loci. We are applying our chromatin catalog to emerging genetic findings of psychiatric disorders and to the identification of developmentally relevant chromatin interactions. Discussion We used eHi-C to generate high-resolution chromatin interaction data from human brain, at two developmental time points, to facilitate the interpretation and prioritization of genetic findings from GWAS. Large-scale connection of GWAS loci to specific genes could yield “actionable” findings, which may lead to the development of more targeted therapies.