Chromatin and gene-regulatory dynamics of human pulmogenesis by single cell multiomic sequencing

Abstract

Human lung development is governed by complex gene regulatory networks that orchestrate cellular differentiation and organogenesis. We present a single cell multiomic atlas of human pulmogenesis, simultaneously capturing both the chromatin accessibility profile and the transcriptome from each cell across fetal lungs spanning from post-conception weeks (PCW) 12 to 23. We identified 44 distinct developing cell clusters and mapped 581,745 candidate cis-regulatory elements and nominated 121,486 non-redundant peak-to-gene linkages. We identify highly regulated genes (HRGs) and the cognate highly regulating peaks (HRPs) that describe the most salient regulatory gene programs and developmental enhancer sites for each cell type. Trajectory analysis along with interpretable cell type specific convolutional neural network models were developed to delineate dynamic regulatory programs driving key developmental transitions, including aerocyte and arterial differentiation and alveolar formation. Furthermore, we identified distinct vascular smooth muscle subpopulations with unique spatial associations to either arterial or venous structures with reciprocal signaling within each niche. We also uncovered the regulatory modules of surfactant production in alveolar progenitors, implicating a direct role for the glucocorticoid receptor alongside novel transcription factors. Finally, using cell type specific models linking DNA sequence to chromatin accessibility we prioritize variants associated with impaired pulmonary function or disease and nominate mechanisms of motif disruption. Overall, our multiomic atlas deepens our understanding of the gene-regulatory architecture underlying human lung development and provides a valuable resource for the community to dissect the cellular and molecular programs of pulmonary physiology and disease at the cellular and nucleotide precision.