CellCover defines marker gene panels capturing developmental progression in neocortical neural stem cell identity

Area of Science:

• Genomics
• Computational Biology
• Cell Biology

Background:

• Accurate cell class definition is crucial for single-cell RNA sequencing (scRNA-seq) atlases.
• Traditional methods identify marker genes individually, overlooking joint gene expression patterns.

Purpose of the Study:

• To develop a novel method for identifying discriminating gene panels for cell type classification.
• To improve upon traditional differential expression (DE) methods by considering genes collectively.

Main Methods:

• Framed marker-panel selection as a combinatorial optimization problem (minimal set-covering).
• Developed CellCover, a computational method to efficiently search for gene panels in binarized scRNA-seq data.
• Validated using blood and brain datasets, and transfer-learning across species.

Main Results:

• CellCover effectively reduces gene redundancy and identifies cell-class-specific signals missed by DE methods.
• Demonstrated conserved cell class identification in neocortical neurogenesis across species.
• Tracked developmental progression in progenitors and neurons using gene panels.

Conclusions:

• CellCover offers a robust and efficient approach to cell type definition in large-scale scRNA-seq studies.
• The method highlights the evolutionary origins of cell-type-specific transcriptomic features, such as those in outer radial glia.