Resolving the link between temporal gene expression dynamics and cell fate decisions is essential to understand organismal development. Developmental trajectories can be reconstructed from single-cell RNA sequencing (scRNA-seq) data, but some lineage branches can be ambiguous or misleading. Now, a study in Science has leveraged the invariant cell lineage of Caenorhabditis elegans to generate a lineage-resolved single-cell atlas of C. elegans embryogenesis.

Credit: Photo Library/Alamy

scRNA-seq data were generated for 86,024 single cells from developing C. elegans embryos. The developmental stage (or ‘embryo time’) of each cell was estimated by comparing its expression profile to time-resolved whole-embryo RNA-seq data. The single-cell data were then visualized with the uniform manifold approximation and projection (UMAP) algorithm and clustered using the Louvrain algorithm. Clusters were subsequently annotated for cell type based on marker gene expression. Recursive generation of ‘sub-UMAPs’ for subsets of cells stratified by cell type or embryo time enabled better resolution of cell or lineage identities. Overall, 93% of cells were assigned a cell type or cell lineage, and 1,068 of the 1,228 branches of the C. elegans embryonic lineage that do not lead to programmed cell death were represented in the final atlas.

Analysis of bifurcating trajectories, in which a parent cell gives rise to two different daughter cell types, suggests that each branchpoint decision hinges upon daughter-specific expression of 3–4 transcription factors (TFs). Interestingly, multi-lineage priming, in which both sets of daughter-specific TFs are expressed by the parent cell, seems to be common and pervasive during C. elegans embryogenesis.

Other UMAP trajectories would not have been easily resolved based solely on single-cell transcriptome data and highlight the limitations of such approaches. Examples include convergent trajectories, in which the gene expression patterns of cell lineages with similar or identical terminal cell fates converge; and discontinuous trajectories, in which the gene expression profile changes abruptly at the final cell division when the progenitor cell becomes terminally differentiated. Taken together, these findings indicate that scRNA-seq data alone are unlikely to be sufficient to resolve developmental trajectories for other organisms: integration with lineage tracing methods and improved computational tools will be required.

scRNA-seq data alone are unlikely to be sufficient to resolve developmental trajectories

Moving forwards, data sets for missing developmental stages will be needed to complete the C. elegans atlas. However, integrated multi-omic analyses of mutant embryos are likely to be required to elucidate the molecular mechanisms underlying cell fate decisions.