CellCover defines marker gene panels capturing developmental progression in neocortical neural stem cell identity
Defining cell classes is central to the analysis of growing single-cell RNA sequencing (scRNA-seq) atlases. Marker genes are most often identified by differential expression (DE) methods that assess genes one at a time, ignoring the redundancy and complementarity revealed when genes are considered jointly. Working with binarized expression data, we instead seek discriminating panels of genes that…
CellCover is a method that defines marker gene panels capable of capturing the developmental progression of neocortical neural stem cell identity. Traditional approaches to identifying cell classes in single-cell RNA sequencing (scRNA-seq) atlases often rely on differential expression (DE) methods, which evaluate genes individually, disregarding the redundancy and complementarity that emerge when genes are examined together.
CellCover, however, treats marker-panel selection as a variant of the minimal set-covering problem in combinatorial optimization. This approach allows for the efficient exploration of a vast space of candidate panels, taking advantage of the large numbers of cells typically present in scRNA-seq data and being robust to zero-inflation.
The authors tested CellCover using blood and brain data, demonstrating that it successfully reduces gene redundancy and uncovers cell-class-specific signals that are not captured by DE methods. Furthermore, transfer-learning experiments across mouse, primate, and human data revealed that CellCover is able to identify conserved cell classes in neocortical neurogenesis and track developmental progression in progenitors and neurons.
The study also examined outer radial glia markers across mammals, revealing that the transcriptomic elements of this crucial cell type may have originated in rodent gliogenic precursors prior to the complete program being established in the primate lineage.
Written by urgent.news from eLife's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
