Abstract
Most human genetic studies assume that connections between genomic regions are adequately captured by linkage disequilibrium (LD). LD reflects historical events that shaped the individuals in a sample, and LD patterns are consistent in DNA sampled from essentially any cell at any point in development from embryonic stem cells to highly specialized post-mitotic neurons. LD-based studies have been highly useful in identifying genetic associations for idiopathic maladies like schizophrenia; however, even as LD is fundamental to these successes, LD confounds attempts to derive confident biological insights from genetic results as it imperfectly reflects the functional state of DNA in cell nuclei. Interpretation of most genetic findings for schizophrenia is complicated by the presence of many significant and highly correlated associations in non-coding regions. Here, we created a high-resolution map of three-dimensional genome architecture by applying Hi-C to adult and fetal brain cortex with concomitant RNA-seq, ATAC-seq, and ChIP-seq data (CTCF, H3K27ac, and H3K4me3). An extensive set of analyses established the quality, information content, and salience of these new Hi-C data. We used these 3D data to functionally connect genetic results for schizophrenia to specific genes. We show that LD-based approaches provide a limited view of the complexity of schizophrenia GWAS findings. Gene set analyses based on functional genomic data provide an expanded view of the biological processes involved in the etiology of schizophrenia.