SCC

SCC imputes dropout events in single-cell RNA sequencing (scRNA-seq) data using a mixture model and nearest-neighbor information to improve gene expression estimates for downstream analyses.

Key Features:

• Nearest Neighbor Identification: Identifies nearest-neighbor cells for each cell to provide local context for imputation.
• Mixture Model Imputation: Fits a mixture model to estimate dropout probabilities and impute gene expression values from sparse scRNA-seq measurements.
• Reduction of Intra-Class Distance: Decreases intra-class distance among cells after imputation, increasing within-class similarity.
• Improved Clustering Accuracy: Improves clustering accuracy relative to existing methods, as observed on simulated and real datasets.

Scientific Applications:

• Dynamic Developmental Processes: Enables analysis of temporal and developmental cell-state transitions by providing more complete expression profiles.
• Gene Regulation Mechanisms: Facilitates study of gene regulation and regulatory networks by reducing dropout-induced noise in expression data.
• Discovery of New Cell Types: Aids identification of novel or rare cell types that may be obscured by dropout noise in raw scRNA-seq data.

Methodology:

Identifies nearest-neighbor cells for each cell and applies a mixture model to estimate dropout probabilities and impute gene expression values.