2DImpute

2DImpute imputes dropout events in single-cell RNA sequencing (scRNA-seq) data by predicting false zero expression values using gene–cell expression relationships.

Key Features:

• Dropout Imputation: Predicts and corrects false zero expression values in scRNA-seq datasets to recover undetected transcripts.
• Gene–Cell Relationship Modeling: Utilizes interdependencies between genes and cells in the expression matrix to guide imputation.
• Flexible Data Input: Accepts normalized scRNA-seq expression matrices in both log-transformed and non-log-transformed formats while preserving matrix dimensions.
• Cross-Protocol Performance: Demonstrates improved imputation accuracy across multiple scRNA-seq experimental protocols.

Scientific Applications:

• Cellular Heterogeneity Analysis: Improves detection of cell subtypes and cellular states by correcting dropout-related expression artifacts.
• Gene Expression Profiling: Enhances accuracy of gene expression measurements for downstream analyses such as differential expression and pathway enrichment.
• Comparative scRNA-seq Studies: Enables consistent analysis of datasets generated from different single-cell RNA sequencing protocols.

Methodology:

The method analyzes a gene-by-cell expression matrix to identify false zero values and imputes them using a predictive model based on gene expression patterns and relationships among cells.