Overview

Animals are composed of cell types that express different gene sets, representing overlapping but different partitions of open chromatin. The logic however of how differentially accessible chromatin across the cells of the body implements their diverging differentiation trajectories is poorly understood. The marine annelid Platynereis dumerilii (Pdum) is ideally suited to address this problem, as it differentiates hundreds of cell types at early young worm stages when the overall number of cells is still low. This implies that we can resolve and compare differentiation trajectories for all cell types at once and across the entire body.
The transcriptional state of differentiating cell types results from underlying gene regulatory networks (GRNs), with limited numbers of transcription factors (TFs) regulating each other and downstream target genes. To identify the TFs and gene regulatory networks driving cell type differentiation, we will use novel multi-omics approaches to identify open chromatin regions for larval and young worm stages across all cells of the body. We will then use machine learning to identify TF binding motifs that are enriched in the regulatory regions that drive the expression of cell type-specific genes, and use this information to generate cell type- and stage-specific GRNs. This will allow us to identify key TFs driving cellular differentiation in all Pdum cell types, and help elucidate general principles of GRNS driving cell type development and evolution

Project

Recent progress in single-cell RNA and ATAC sequencing provides new opportunities for the high-resolution generation of GRNs during the differentiation of specific cell types and tissues 1,2. For rare cell types, however, GRN reconstruction remains a particular challenge due to the relative scarcity of DNA. Also, in conventional models analysis is limited to restricted tissues due to exorbitant cell numbers, which precludes a whole-body view of GRN analysis and hinders comparison between remote species with non-corresponding tissue composition. We thus decided to focus our efforts on the dissection of GRNs driving differentiation in the marine annelid Pdum, an emerging model for development and evolution, neurobiology and ecology3. In this model, the bulk of differentiated cell types develops already at the young worm stage that comprises no more than 12,000 cells, for which a unique atlas combining cellular-resolution gene expression and morphology has recently been created4. This is possible because up to this stage, development is highly synchronous and stereotypical, so that thousands of similarly-staged individuals can be easily obtained from one batch of siblings derived from the same parental cross, thus relatively enriching for rare cell types. Also, the Pdum genome is meanwhile available in high quality and chromosome-resolution (Nzumbi Mutemi et al, in prep.) and transgenic approaches are available5. However, since TF binding is only partially conserved across evolution, cis-reguatory motifs are difficult to identify in Pdum and the peculiarities of its regulatory grammar are yet unknown. Our aim is to combine the unique expertises and strengths of our laboratories to enable high-end multi-omics approaches in Pdum, and to use deep learning models to de novo-identify TF binding motifs in the Pdum open chromatin. We will then infer GRNs for the differentiation trajectories leading to all Platynereis cell types and thus obtain the first development and differentiation- GRN with cellular resolution for an entire body. This will be a powerful reference for the comparison and integration with similar datasets recently obtained for flatworms (Bo Wang et al., in prep.) and tissue-restricted ATAC-seq mammalian resources.

This project is supported by a Data Creation Fund

Find out more:

Interested in finding out more about the pharmacogenomics project? Do you want to join our researchers? Get in touch, we would love to hear from you!