Analysis of Technical and Biological Variability in Single-Cell RNA Sequencing

Area of Science:

• Genomics
• Computational Biology
• Molecular Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) enables detailed characterization of cellular heterogeneity.
• Technical variability in scRNA-seq gene expression is significantly higher than in bulk RNA-seq.
• Accurate analysis of scRNA-seq data necessitates accounting for technical variability.

Purpose of the Study:

• To describe a computational pipeline for detecting highly variable genes in scRNA-seq data.
• To identify genes with cell-to-cell variability exceeding technical noise.
• To provide a framework for robust analysis of single-cell transcriptomic data.

Main Methods:

• Utilized the scater and scran R/Bioconductor packages.
• Implemented deconvolution-based normalization.
• Applied mean-variance trend fitting and statistical testing for biological variability.
• Included visualization of highly variable genes.

Main Results:

• Developed and presented a computational pipeline for identifying highly variable genes.
• Demonstrated the pipeline's effectiveness using mouse embryonic stem cells and dentate gyrus cells.
• Provided a method to distinguish biological variability from technical noise in scRNA-seq.

Conclusions:

• The presented pipeline effectively detects highly variable genes in scRNA-seq data.
• Accounting for technical noise is essential for reliable single-cell gene expression analysis.
• This approach enhances the understanding of cellular heterogeneity through accurate gene variability assessment.