Vireo: Bayesian demultiplexing of pooled single-cell RNA-seq data without genotype reference.

Area of Science:

• Genomics
• Computational Biology
• Molecular Biology

Background:

• Multiplexed single-cell RNA sequencing (scRNA-seq) enables higher throughput and batch effect correction.
• Demultiplexing cells to their original samples typically requires complete genotype data, limiting its application.
• Existing methods are constrained by the need for comprehensive genetic information.

Purpose of the Study:

• To introduce Vireo, a computationally efficient Bayesian model for demultiplexing scRNA-seq data from pooled samples.
• To enable demultiplexing even with partial or no prior genotype information.
• To broaden the applicability of multiplexed scRNA-seq experimental designs.

Main Methods:

• Development of a Bayesian model named Vireo.
• Application of Vireo to synthetic mixtures and real scRNA-seq data.
• Utilizing genetic variants as natural barcodes for cell demultiplexing.

Main Results:

• Vireo demonstrates robust performance in demultiplexing pooled scRNA-seq data.
• The model successfully demultiplexes cells without requiring complete genotype information.
• Validation using both synthetic and real experimental datasets.

Conclusions:

• Vireo enhances the utility of multiplexed scRNA-seq designs by overcoming genotype data limitations.
• The method provides a flexible and efficient approach for cell demultiplexing.
• Facilitates broader adoption of pooled scRNA-seq for various expression analyses.