What are the molecular trajectories that embryonic cells go through as they become specialized? In a study in Science (PDF) the Schier lab combines single-cell RNA-sequencing with a novel computational approach to generate a branching tree that describes the molecular specification of 25 different cell types in the zebrafish embryo. Analysis of the tree reveals how cells acquire their fates.
All animals begin as a single, fertilized cell, which means this single cell must both proliferate and differentiate into all distinct types of cells in order to make a functional animal. Understanding how this diversity of cell types arises during development is important both for answering the basic question of how organs and animals build themselves, and also for instructing efforts to engineer these processes to study and treat disease.
The genetic control of cell type specification during development has been investigated for decades, but classic studies in the field have had to focus on particular cell fate decisions and a handful of genes at a time. Recent technological advances have enabled high-throughput single-cell RNA sequencing, where the complement of genes activated in individual cells can be profiled. This approach allows the gene expression profiles of embryonic cells to be captured as they acquire their fates. Applying this technology to investigate zebrafish development, Jeffrey Farrell and Yiqun Wang from the Schier lab, with assistance from Samantha Riesenfeld and Karthik Shekhar in Aviv Regev’s lab at the Broad Institute, profiled over 38,000 zebrafish embryonic cells from 12 stages spanning 9 hours of development. This period of time begins just after embryonic cells begin to show differences in their gene expression and ends when dozens of distinct cell types can be recognized by their morphologies or expression of distinct marker genes.
The authors then developed a computational technique, URD, that looked for cells that had very similar gene expression and used those connections to uncover the paths through gene expression that cells take as they adopt their specific fates. Their approach generated a branching tree that described the molecular specification of 25 different cell types in the zebrafish embryo. Study of the branching tree revealed how cells change their gene expression as they become specialized, recovering both classic and new markers of cell populations, and suggesting candidate regulators of cell specification events. By combining the branching tree with spatial inference, the authors also reconstructed the spatial origins in the early embryo of different cell types present at the end of their timecourse. Additionally, by profiling cells from a mutant that lacks a developmental signal at an early timepoint, the authors showed that mutant cells adopted a subset of wild-type gene expression states, while no new cell states were found. Furthermore, using their reconstructed branching tree, they were able to predict the tissues that would later be missing in the mutant. Finally, some developmental branchpoints contained intermediate cells that expressed genes characteristic of multiple downstream cell fates. Further analysis of one such branchpoint suggested that the intermediate cells switched their specification from one fate to another.
Overall, the scRNA-seq data and developmental tree generated in this study provide a rich resource for future studies of embryonic development. Moreover, the approach provides a broadly applicable framework to reconstruct complex developmental trajectories from egg to adult, and can be potentially applied to study other processes, such as the progression of disease.