Virtual tissue expression analysis.

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Bulk RNA sequencing data is abundant but lacks cellular resolution.
• Single-cell RNA sequencing provides detailed cellular insights but is less accessible.
• Distinguishing cell type-specific gene expression from bulk data is a significant challenge.

Purpose of the Study:

• To develop and validate an algorithm, tissueResolver, for deconvoluting single-cell expression profiles from bulk RNA sequencing data.
• To enable the attribution of gene expression changes to specific cell types within a tissue sample.
• To apply tissueResolver to uncover cell type-specific regulatory differences in diffuse large B-cell lymphoma subtypes.

Main Methods:

• Development of the tissueResolver algorithm for single-cell data extraction from bulk RNA-seq.
• Validation of tissueResolver using simulated datasets to compare performance against existing methods.
• Application of tissueResolver to analyze gene expression patterns in activated B-cell-like (ABC) and germinal center B-cell-like (GCB) DLBCL subtypes.

Main Results:

• tissueResolver demonstrates superior performance compared to competing methods on simulated data.
• The algorithm successfully deconvolutes single-cell information from bulk RNA sequencing data.
• Distinct cell type-specific regulatory patterns were identified between ABC and GCB DLBCL subtypes using tissueResolver.

Conclusions:

• tissueResolver is an effective tool for inferring single-cell resolution from bulk RNA sequencing data.
• The algorithm provides valuable insights into cellular heterogeneity and cell type-specific gene regulation.
• tissueResolver facilitates the discovery of molecular differences between disease subtypes at the single-cell level.