DSB

DSB normalizes and denoises protein expression counts from droplet-based CITE-seq experiments to improve single-cell proteomic quantification and is implemented as an R package.

Key Features:

• Normalization: Adjusts CITE-seq protein counts to account for technical variability and enable comparability across cells.
• Denoising: Reduces background signal and technical noise to enhance protein signal-to-noise ratio.
• Protein-specific background correction: Models and mitigates background arising from unbound oligo-conjugated antibodies.
• Cell-specific noise estimation: Captures and corrects cell-level variability using shared variance among isotype controls and background protein counts.

Scientific Applications:

• Cellular heterogeneity analysis: Improves quantification of protein markers for dissecting cell populations at single-cell resolution.
• Clustering and cell-type identification: Enhances downstream clustering by reducing technical noise in protein measurements.
• Differential protein expression: Enables more accurate testing for protein-level differences between conditions or cell types.
• Multi-omics integration: Facilitates integration of CITE-seq protein data with transcriptomic and other omics datasets for joint analyses.
• Biomarker discovery: Supports identification of candidate protein biomarkers by improving data quality.

Methodology:

Models and subtracts protein-specific background from oligo-conjugated antibodies, estimates cell-specific noise via shared variance among isotype controls and background protein counts, and applies normalization to CITE-seq protein counts.