SCRABBLE

SCRABBLE imputes dropout values in single-cell RNA sequencing (scRNA-seq) data by integrating bulk RNA-seq constraints to reduce imputation bias and better preserve gene-gene and cell-cell relationships.

Key Features:

• Enhanced Dropout Recovery: Demonstrates improved identification and recovery of dropout events in scRNA-seq data compared to existing methods.
• True Gene Expression Distribution: Captures authentic gene expression distributions across individual cells to reflect cellular heterogeneity.
• Preservation of Relationships: Maintains fidelity of gene-gene and cell-cell relationships within imputed scRNA-seq datasets.
• Bias Reduction via Bulk RNA-seq Constraint: Incorporates bulk RNA-seq data as a constraint to reduce bias toward expressed genes during imputation.

Scientific Applications:

• Developmental Biology: Improves resolution of cellular differentiation trajectories by reducing dropout-related distortions in scRNA-seq data.
• Cancer Research: Enhances detection of tumor heterogeneity and rare cell populations by providing more accurate single-cell expression estimates.
• Immunology: Refines characterization of immune cell subsets and states through more reliable single-cell transcriptomes.
• Single-cell Transcriptomics Quality Improvement: Serves as a preprocessing step to improve downstream analyses that rely on accurate scRNA-seq expression values.

Methodology:

Integrates bulk RNA-seq as a constraint within an imputation algorithm and evaluates performance using simulated and experimental scRNA-seq data.