[Imputation method for dropout in single-cell transcriptome data].

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Context:

• Single-cell RNA sequencing (scRNA-seq) offers high-resolution insights into cellular heterogeneity and tissue dynamics.
• Technical zeros, arising from limitations in scRNA-seq, pose a significant challenge for accurate downstream analyses.
• These zeros can obscure true biological signals, impacting cell clustering, differential gene expression, and pseudotime analysis.

Purpose:

• To review and critically assess existing methods for imputing technical zeros in scRNA-seq data.
• To highlight the advantages and disadvantages of various imputation techniques.
• To provide recommendations and future perspectives for the development and application of these methods.

Summary:

• This paper surveys computational approaches designed to address technical zeros in scRNA-seq data.
• The core strategy involves leveraging correlations between cells and genes to infer missing expression values.
• The review discusses the strengths and weaknesses of current imputation algorithms.

Impact:

• Improved accuracy in downstream analyses such as cell clustering and differential gene expression.
• Enhanced ability to identify subtle biological signals and cell populations.
• Facilitation of more robust and reliable biological discoveries from scRNA-seq experiments.