Imputing missing values in single-cell RNA-sequencing data: a statistical and machine learning-based approach

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) is crucial for understanding cellular heterogeneity.
• Dropout events, characterized by excessive zero counts, are a significant challenge in scRNA-seq data.
• These dropouts can obscure true biological signals and hinder downstream analyses.

Purpose of the Study:

• To develop a novel computational method for detecting and imputing dropout events in scRNA-seq data.
• To improve the accuracy of gene expression quantification and downstream analysis of scRNA-seq data.
• To provide a robust tool for addressing data sparsity in single-cell studies.

Main Methods:

• Proposed single-cell dropout detection and imputation (scDDI) method.
• Utilized a Poisson-negative binomial mixture model for dropout event identification.
• Employed a decision tree regression model for imputing missing gene expression values.

Main Results:

• scDDI demonstrated superior performance in dropout detection compared to existing methods.
• The method effectively imputed missing values in both simulated and real scRNA-seq datasets.
• scDDI significantly enhanced the performance of downstream tasks, including gene expression recovery, cell clustering, and subpopulation identification.

Conclusions:

• scDDI offers a powerful solution for addressing data sparsity and dropout events in scRNA-seq.
• The method improves the reliability and accuracy of single-cell gene expression analysis.
• scDDI facilitates more robust cell subpopulation identification and biological discovery from scRNA-seq data.