Bridging morphogenesis and single cell transcriptomics during C. elegans early development
Metazoan develop from a single fertilized egg. This process involves the spatio-temporal unfolding of the cell lineage tree into an organized shape with functional parts. The sequence of events leading to the precise positioning of individual cells with the required fate is orchestrated by gene regulation within and between cells, cell proliferation and rearrangements as well as global morphological changes. Even though various parts of these processes have been elucidated, a generic framework that would capture the interactions between them is still missing. We propose to describe a developing organism as an evolving graph where each cell correspond to a node, spatial relationships are encoded as the edges and transcriptomic state as label on the nodes. We will explore how the action of cell proliferation, spatial rearrangement and differentiation dynamics affect the characteristics of these graphs using published data and modeling in the development of the worm C. elegans.
Graph theory, single cell RNASeq, UMAP, manifold learning, optimal transport
This project aims at combining the study of morphogenesis with the study of cell differentiation from a computational perspective. Two main questions will be answered. First, given the constraints of cell proliferation, cell spatial ordering and cell-cell signaling, can we show that the process of cell differentiation and specification is optimal at a global level? Second, given a shape and a functional organization, can we predict the cell proliferation and differentiation process that gave rise to it? And if not, where do the limitations come from?
Proposed approach (experimental / theoretical / computational)
The proposed approaches will be based on graph theory and manifold learning techniques. Cell detection in developing embryos enables the construction of graphs, where nodes are the cells and edges encode for their contact information. These graphs evolve over time and can be compared to the cell lineage. We will use the tools of graph theory to analyze the relationships between cell proliferation and spatial organization. Alongside, single cell RNASeq data gives rise to point clouds in high dimensional spaces. We will study their shape using manifold learning techniques and compare transcriptomic states of different cells. This leads to a graph where the edges encode transcriptional similarities between cells. In addition, spatial relationships of cells in C. elegans will be examined through physical modeling of cell-cell contacts during proliferation in a finite volume, following up on an ongoing collaboration with Charlotte Rulquin (postdoc in PF Lenne’s lab). Finally, this work will contribute to the open question of tracing the origin and extent of left-right functional asymmetry in C. elegans nervous system examined in the Bertrand laboratory. We will design new experiments from prediction obtained by this computational approach through a related PhD project (“An interdisciplinary approach to the study of lateralization in the nervous system” https://centuri-livingsystems.org/phd2020-20/). Although benefiting highly from this joint project, the proposed computational approach stands by itself and the increased demand of new methods providing meaningful predictions from single cell RNASeq data makes it very relevant for various areas of developmental biology.
The variety of methods proposed to tackle the problem of how to bridge morphogenesis with the study of cell differentiation at single cell resolution clearly lead to a highly interdisciplinary project. The approaches that we will be using come from the fields of applied mathematics (graph theory and high dimensional statistics), physics (modeling and simulation of a multiagent system) and biology. The possibility of initiating new experiments driven by questions emerging from the computational aspects of the project show the interests that the project can have in biology. The system C. elegans has been chosen for its simplicity (rapid and transparent development, invariant cell lineage), however, the results of this project are expected to have a broad impact in the field of developmental biology in general. On the other hand, the methods developed require an expert knowledge in graph theory and manifold learning techniques. The specificity of studying an evolving graph observed in multiple spaces will lead to new theoretical and methodological questions that go beyond biology.
We are looking for a student with background in physics or applied mathematics with strong interests in developmental biology. The project is mainly computational and will be carried out in P. Villoutreix’s group with extensive support from the Bertrand Lab. Fruitful interactions are expected with an experimental biologist student recruited in the team of V. Bertrand who will develop biological experiments on a complementary project but the two recruitments are independent (see project “An interdisciplinary approach to the study of lateralization in the nervous system” https://centuri-livingsystems.org/phd2020-20/).