Abstract 349: Gene Regulatory Modules Regulate Cardiac Lineage Specification Revealed by Single-cell Rna Sequencing

Abstract

Background: The molecular definition of the differentiation process is of intense interest for cardiovascular developmental biologists. With the advent of single cell technologies, a number of a transcriptomic approaches have been developed to obtain a complete description of gene expression. While single cell RNA sequencing is a significant advance for describing gene expression at the cellular resolution, novel obstacles have emerged such as computational management of dropout events, discovery of the gene regulatory modules responsible for lineage specification and identification of novel lineage markers. Methods: Using Etv2- and Nkx2-5 EYFP transgenic embryos, from which progenitors of hematopoietic, endothelial and cardiac lineages can be isolated with high purity, we performed single cell transcriptome analysis. We developed a Bayesian statistical framework to (1) model the dropout events, (2) reconstruct the lineage specification pathways from mesodermal progenitors to hematopoietic, endothelial and cardiac lineages, (3) discover the core gene regulatory modules that govern lineage specification and (4) predict novel cell surface markers for cardiac components. We developed a structural stochastic variational inference method to efficiently estimate the model parameters for sequencing data with thousands of cells. Results: The analysis of 421 Etv2- and Nkx2-5 EYFP positive cells predicted the core regulatory networks involving transcription factors, signaling molecules and cell surface proteins, where the signaling cascades are significantly correlated the specification of the endothelial, hematopoietic and various cardiac lineages from the mesodermal progenitors. The method also predicted novel cell surface markers for the first and second heart fields. Conclusions: We report the analysis of complete gene expression profiles of mesodermal lineage specification at single cell resolution. We provide the first unified statistical framework to systematically analyze single cell RNA-seq data and quantitatively couple the network analysis and lineage specification. This strategy provides a platform for the identification of regulatory nodes and cascades that govern mesodermal and cardiovascular development.