Sources of variation in cell-type RNA-Seq profiles

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• Accurate cell-type specific gene expression profiles are essential for computational analyses of bulk RNA-Seq data, including deconvolution and digital cytometry.
• Variations in these profiles due to technical factors and biological differences can reduce the efficacy of these methods.

Purpose of the Study:

• To investigate the primary factors contributing to variation in cell-type specific gene expression profiles.
• To evaluate the impact of these variations on computational methods like deconvolution.
• To compare UMI-based single-cell RNA-Seq with bulk RNA-Seq.

Main Methods:

• Evaluation of various normalization methods.
• Quantification of variance explained by different factors (e.g., laboratory, tissue of origin).
• Analysis of publicly available bulk and single-cell RNA-Seq datasets containing B and T cells.

Main Results:

• Substantial technical variation exists across laboratories, even for genes critical for deconvolution, confounding results.
• Tissue of origin is a significant factor, posing challenges when using blood-derived profiles for other tissues.
• Differences between UMI-based single-cell and bulk RNA-Seq are largely attributable to read duplicate counts per mRNA molecule.

Conclusions:

• Technical factors, particularly laboratory variation and tissue of origin, substantially influence cell-type specific gene expression profiles.
• These variations confound computational analyses, emphasizing the need for careful consideration in profile generation.
• Matching or correcting for technical factors is vital when creating cell-type specific profiles for use with bulk RNA-Seq samples.