Single-nucleus ATAC-seq elucidates major modules of gene regulation in the development of non-alcoholic fatty liver disease

Abstract Background Non-alcoholic fatty liver disease (NAFLD) develops from fatty liver to steatohepatitis during which multiple cell types may play different roles. Aiming to understand tissue composition of cell types, their gene expression and global gene regulation in the development of NAFLD, we performed single-nucleus and bulk ATAC-seq on the liver of rats fed with a high-fat diet (HFD). Methods Male Spontaneously Hypertensive Rats were fed a normal diet or a HFD. Rats fed HFD for 4 weeks developed fatty liver, and those fed HFD for 8 weeks further developed steatohepatitis. Under the washout condition, where 4 weeks of HFD is followed by 4 weeks of a normal diet, fatty liver was partially ameliorated. For each dietary condition, we performed single-nucleus ATAC-seq on one animal and bulk ATAC-seq on four animals. Results In accordance with the pathological progression from fatty liver to steatohepatitis, the proportion of inflammatory macrophages dramatically increased. By machine learning, we divided global gene expression into modules, such that transcription factors in a module regulate a set of genes in the same module. Consequently, many of the modules rediscovered known regulatory relationship between the transcription factors and biological processes. For the discovered biological processes, we searched core genes, which were defined as genes central regarding co-expression and protein-protein interaction. A large part of the core genes overlapped with previously implicated NAFLD genes. Conclusions Single-nucleus ATAC-seq combined with data-driven statistical analysis help elucidate the global gene regulation in vivo as a combination of modules and discover core genes of the relevant biological processes.