Integrating single-cell transcriptomic data across different conditions, technologies, and species.

Area of Science:

• Computational Biology
• Genomics
• Single-cell Analysis

Background:

• Single-cell RNA sequencing (scRNA-seq) excels at defining cellular phenotypes within single experiments.
• Integrating multiple scRNA-seq datasets to identify shared cell subpopulations across diverse conditions remains a significant challenge.
• Existing methods struggle with cross-dataset comparisons, limiting comprehensive biological understanding.

Purpose of the Study:

• To introduce a novel analytical strategy for integrating scRNA-seq data from multiple sources.
• To enable the identification of shared and distinct cellular subpopulations across datasets.
• To facilitate downstream comparative analyses for deeper biological insights.

Main Methods:

• Developed an analytical strategy for scRNA-seq data integration based on common sources of variation.
• Implemented the strategy in the R toolkit Seurat (http://satijalab.org/seurat/).
• Applied the approach to diverse datasets: peripheral blood mononuclear cells (resting/stimulated), hematopoietic progenitors (two technologies), and human/mouse pancreatic islet cell atlases.

Main Results:

• Successfully aligned and integrated multiple scRNA-seq datasets, revealing shared and transitional cell states.
• Demonstrated the ability to identify common cell populations across different experimental conditions, technologies, and species.
• Enhanced statistical power through integrated analysis, enabling robust discovery of cell states.

Conclusions:

• The proposed analytical strategy effectively integrates scRNA-seq data, overcoming previous limitations.
• This approach facilitates robust comparisons across datasets, improving the identification of shared cell populations.
• Enables deeper understanding of cellular responses to perturbations, disease, and evolutionary processes.