CosGeneGate selects multi-functional and credible biomarkers for single-cell analysis
Tianyu Liu1,2, Wenxin Long1,3, Zhiyuan Cao1,2,4
1Department of Biostatistics, Yale University, New Haven, CT, 06520, United States.
Briefings in Bioinformatics
|November 26, 2024
Summary
CosGeneGate effectively selects non-redundant marker genes for cell type identification in single-cell sequencing. This novel model improves cell-type classification accuracy and downstream analysis performance over existing methods.
Area of Science:
- Genomics
- Computational Biology
- Bioinformatics
Background:
- Marker gene selection is crucial for cell type identification in single-cell sequencing.
- Existing methods often yield redundant genes or genes lacking cell-type specificity.
Purpose of the Study:
- To introduce CosGeneGate, a novel model for selecting effective marker genes.
- To enhance cell type classification accuracy and specificity in single-cell analysis.
Main Methods:
- Developed CosGeneGate, a model integrating cell-type classification accuracy and marker gene expression specificity.
- Evaluated CosGeneGate using public and newly generated single-cell datasets.
Main Results:
- CosGeneGate outperforms existing methods in marker gene selection.
- Selected marker genes demonstrate superior performance in downstream analyses.
- Identified non-redundant marker genes for major human cell types and tissues.
Conclusions:
- CosGeneGate offers a more effective approach to marker gene selection for single-cell sequencing.
- The identified marker genes facilitate improved cell type resolution and analysis.
