Deconvolution from bulk gene expression by leveraging sample-wise and gene-wise similarities and single-cell RNA-Seq...

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• Bulk RNA sequencing (RNA-seq) averages gene expression, obscuring crucial cell type heterogeneity.
• Single-cell RNA-seq (scRNA-seq) reveals cellular heterogeneity but is resource-intensive and impractical for large-scale studies.
• Existing computational deconvolution methods often require prior knowledge of either cell type composition or cell type-specific gene expression profiles.

Purpose of the Study:

• To develop a novel computational deconvolution algorithm for simultaneous inference of cell type proportions and cell type-specific gene expression profiles (GEPs) from bulk RNA-seq data.
• To provide a practical and accurate alternative to experimental methods for characterizing cellular heterogeneity.

Main Methods:

• Developed DSSC, a deconvolution algorithm that leverages gene-gene and sample-sample similarities in bulk RNA-seq and scRNA-seq data.
• Evaluated DSSC's performance on simulated pseudo-bulk data (intra- and inter-dataset) and experimental bulk data (mixture and real samples).

Main Results:

• DSSC accurately infers both cell type proportions and cell type-specific GEPs simultaneously.
• The method demonstrates robustness across various simulation types and experimental datasets.
• DSSC exhibits efficiency in terms of cost and time compared to existing approaches.

Conclusions:

• DSSC offers a practical and effective computational solution for dissecting cellular composition and gene expression heterogeneity in bulk samples.
• This method advances the study of biological mechanisms by providing detailed insights into cellular heterogeneity without the need for expensive experimental techniques.