LRcell: detecting the source of differential expression at the sub-cell-type level from bulk RNA-seq data

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Bulk RNA sequencing (RNA-seq) analyzes gene expression in mixed cell populations, obscuring cell-type-specific changes.
• Identifying the specific cell types responsible for differential gene expression in bulk RNA-seq remains a significant challenge.
• Single-cell RNA sequencing (scRNA-seq) can resolve cell-type-specific expression but is resource-intensive.

Purpose of the Study:

• To introduce LRcell, a novel computational method for identifying cell (sub)types driving differential expression in bulk RNA-seq.
• To provide a cost-effective and efficient alternative to scRNA-seq for cell-type-specific expression analysis.
• To generate hypotheses regarding cell-type contributions to observed expression changes in bulk RNA-seq experiments.

Main Methods:

• LRcell utilizes bulk RNA-seq data and optionally leverages pre-computed marker genes from scRNA-seq experiments.
• The method performs a simulation study to validate its effectiveness and reliability.
• LRcell is applied to three distinct real-world datasets, including those related to psychiatric disorders.

Main Results:

• Simulation studies confirm the accuracy and dependability of the LRcell computational method.
• Application to real datasets successfully identified known cell types implicated in psychiatric disorders.
• LRcell effectively generates hypotheses about the specific cell (sub)types contributing to differential gene expression.

Conclusions:

• LRcell is a valuable computational tool for dissecting cell-type-specific gene expression from bulk RNA-seq data.
• The method complements existing cell type deconvolution techniques, offering a new approach for biological interpretation.
• LRcell facilitates hypothesis generation regarding the cellular basis of differential expression in complex tissues.