scDA: Single cell discriminant analysis for single-cell RNA sequencing data.

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• Single-cell RNA-sequencing (scRNA-seq) enables deep investigation of cellular heterogeneity.
• Analyzing large-scale scRNA-seq data presents significant computational challenges for cell population characterization.

Purpose of the Study:

• To introduce single cell discriminant analysis (scDA) for efficient and accurate analysis of large scRNA-seq datasets.
• To address the computational challenges in identifying and annotating diverse cell populations.

Main Methods:

• scDA constructs a cell-by-cell representation graph to identify cell groups and discriminant metagenes.
• The method uses these features to simultaneously group cells and annotate unlabeled cells.

Main Results:

• scDA effectively determines cell types and reveals cellular variabilities across eleven datasets.
• The approach outperforms existing methods in labeling new samples and is robust to dropout events.
• scDA can label numerous cells across datasets after training on limited data.

Conclusions:

• scDA provides an efficient computational approach for analyzing large-scale or multi-batch scRNA-seq data.
• The method enhances the ability to accurately characterize cell populations and annotate cells.