scDoc

scDoc corrects drop-out events in single-cell RNA sequencing (scRNA-seq) data by imputing missing gene expression values using information from highly similar cells to improve downstream analyses.

Key Features:

• Drop-out awareness: Explicitly accounts for drop-out events that arise from extremely low mRNA input or stochastic gene expression variability when analyzing scRNA-seq data.
• Similarity-based estimation: Incorporates direct consideration of drop-outs into the estimation of cell-to-cell similarity.
• Local imputation: Imputes missing gene expression by borrowing information from neighboring cells exhibiting high similarity.
• Preservation of biological variability: Aims to preserve inherent biological variability while correcting technical drop-out artifacts.
• Evaluated performance: Performance has been assessed on both simulated datasets and real-world scRNA-seq studies.

Scientific Applications:

• Data visualization: Improves clarity of dimensionality reduction plots and other visualization outputs for scRNA-seq data.
• Cell subpopulation identification: Enhances accuracy of identifying cell subpopulations and novel cell subtypes.
• Differential expression analysis: Bolsters detection of differentially expressed genes in single-cell studies.
• Cellular heterogeneity and development studies: Supports analyses of cellular heterogeneity and developmental processes at single-cell resolution.

Methodology:

scDoc incorporates drop-out events into cell-to-cell similarity estimation and imputes dropped-out gene expression by borrowing values from highly similar neighboring cells; performance was evaluated on simulated and real scRNA-seq datasets.