Science Inventory

Teratogenesis at the single-cell level: opportunities and challenges


Knudsen, T. Teratogenesis at the single-cell level: opportunities and challenges. Society for Birth Defects Research and Prevention (formerly the Teratology Society) Annual Meeting, Charleston, South Carolina, June 28 - July 01, 2020.


Presentation given to the Society for Birth Defects Research and Prevention (formerly the Teratology Society) annual meeting June 2020. This talk will introduce a symposium on "Single-cell revolution: embryogenesis at high-resolution".


This presentation will provide a general overview of the scientific rationale and main concepts behind single cell RNA profiling (scRNA-seq) for comprehensive interrogation of genomic function. scRNA-seq enables high-resolution cell lineage maps (continuum gene expression manifold) when in combination with computational approaches to biological (sampling variance, cell dissociation, batch effects) and numerical (data integration, clustering, rare cell states/types) challenges. For example, t-distributed stochastic neighbor embedding (t-SNE) or uniform manifold approximation and projection (UMAP) methods reduce the dimensionality of cell-cell distances to virtual ‘pseudo-time’ state trajectories and bifurcation points. In embryology, this methodology has been applied to profile lineage trees across thousands of cells in simple model organisms (e.g., Planaria, C elegans, Drosophila) and embryonic stem cells during differentiation. Such tree-like lineages can decipher critical states during vertebrate development from fertilization-gastrulation-organogenesis (zebrafish, Xenopus, chick, mouse) and mis-specification of subnetworks, such as cardiac progenitor cells in Hand2-null mice or neuroprogenitor cell programming during brain development. Combining scRNA-seq with other single-cell modalities such as light sheet microscopy can further couple transcriptomic state profiles with cell fate and behavior. For birth defects research and prevention, single cell multi-modal approaches can be applied to: (i) infer benchmark dose and time at which altered gene expression begins in a point-of-departure (PoD) where bulk averaging methodology is insensitive to small, sentinel changes; (ii) spatially dissect a target field by similarity in cell-level transcriptome response queried against reference single cell atlases for mouse and human (MCA, HCA); (iii) pseudo-time earmarking of sentinel cells following a physiological stimulus or perturbation; and (iv) reverse-engineering attractor states that determine when a complex dynamical system hits its toxicological tipping point; and (v) virtual reconstruction of cell signaling networks during the critical period following chemical perturbation. This abstract does not represent the views of the Agency.

Record Details:

Product Published Date: 07/01/2020
Record Last Revised: 07/08/2020
OMB Category: Other
Record ID: 349283