scMEB

scMEB identifies differentially expressed genes (DEGs) from single-cell RNA sequencing (scRNA-seq) data without relying on prior cell clustering.

Key Features:

• Clustering-Independent Analysis: Identifies DEGs without requiring prior cell clustering.
• Utilization of Stably Expressed Genes: Uses known non-DEGs (stably expressed genes) as reference points to construct a minimum enclosing ball in a multidimensional feature space.
• Minimum Enclosing Ball Distance Metric: Detects DEGs based on their spatial distance from the center of the hypersphere defined by the minimum enclosing ball.
• High Computational Efficiency: Operates significantly faster than existing non-clustering-dependent methods for high-throughput scRNA-seq datasets.
• Performance on Real Datasets: On 11 real datasets, outperformed other non-clustering-dependent approaches in accuracy of cell clustering, prediction of genes with biological functions, and identification of marker genes.

Scientific Applications:

• Gene Expression Profiling: Facilitates identification of DEGs for single-cell gene expression profiling without preliminary clustering.
• Marker Gene Identification: Pinpoints marker genes to distinguish cell types and states within heterogeneous populations.
• Biological Function Prediction: Supports prediction of gene biological functions by prioritizing genes differentially expressed across conditions or cell types.

Methodology:

Construct a minimum enclosing ball (hypersphere) in a multidimensional feature space using stably expressed genes as reference points, map genes into this space, and identify DEGs based on their distance from the hypersphere center.