Network-Based Single-Cell RNA-Seq Data Imputation Enhances Cell Type Identification.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Single-cell RNA sequencing (scRNA-seq) provides high-resolution transcriptomic data.
• scRNA-seq data is characterized by sparsity and dropout events, hindering downstream analyses.
• Accurate gene expression profiles are crucial for cell type identification and understanding cellular heterogeneity.

Purpose of the Study:

• To introduce netImpute, a novel network-based imputation method for scRNA-seq data.
• To address the challenge of dropout events in scRNA-seq data.
• To improve the accuracy of cell type identification and data visualization.

Main Methods:

• netImpute utilizes gene co-expression networks to impute missing expression values.
• The method employs Random Walk with Restart (RWR) to leverage network information.
• Performance was evaluated on simulated and seven real scRNA-seq datasets.

Main Results:

• netImpute effectively recovers missing transcripts and reduces data sparsity.
• The method significantly enhances clustering accuracy and data visualization clarity.
• Gene co-expression networks proved more beneficial than PPI or cell co-expression networks.

Conclusions:

• netImpute offers an effective solution for the dropout problem in scRNA-seq data.
• The method improves the identification and visualization of heterogeneous cell types.
• netImpute enhances the reliability of scRNA-seq data for biological discovery.