Evaluation of zero counts to better understand the discrepancies between bulk and single-cell RNA-Seq platforms

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• Single-cell RNA sequencing (scRNA-Seq) enables high-resolution transcriptome profiling.
• scRNA-Seq data exhibit higher zero-read counts than bulk RNA-Seq, impacting low-expression gene detection.
• Platform-specific discrepancies exist irrespective of gene expression levels.

Purpose of the Study:

• Investigate technical and biological factors causing expression shifts in scRNA-Seq data.
• Identify analytical methods for cross-platform expression shift correction.
• Compare gene and pathway expression differences between single-cell and bulk RNA-Seq.

Main Methods:

• Utilized four paired datasets with multiple samples.
• Employed two distinct machine learning models to analyze data.
• Assessed factors including RNA integrity, gene/UTR3 length, and transcript counts.

Main Results:

• Expression level, RNA integrity, gene/UTR3 length, and transcript number influence zero-read occurrence.
• Identified 25 genes (0.12%) and 7 pathways (2.02%) consistently discordant between single-cell and bulk levels across datasets.
• Discrepancies observed regardless of gene expression levels.

Conclusions:

• Technical factors significantly contribute to expression shifts in scRNA-Seq.
• Findings support the development of novel analytical methods for cross-platform data normalization.
• Understanding these differences is crucial for accurate interpretation of transcriptomic analyses.