Sincast: a computational framework to predict cell identities in single-cell transcriptomes using bulk atlases as...

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Single-cell RNA sequencing (scRNA-seq) is crucial for cell identity characterization.
• Existing tools face challenges with batch effects and limited reference phenotype data.
• Projecting scRNA-seq data onto bulk reference atlases offers a solution for leveraging rich phenotype information.

Purpose of the Study:

• Introduce Sincast, a novel computational framework for querying scRNA-seq data.
• Enable projection of scRNA-seq data onto bulk reference atlases for cell identity prediction.
• Address challenges of batch effects and data sparsity in scRNA-seq analysis.

Main Methods:

• Developed Sincast for projecting scRNA-seq data onto bulk reference atlases.
• Transformed single-cell data using pseudo-bulk aggregation or graph-based imputation for comparability.
• Predicted cell identity along a continuum to identify novel cell states.

Main Results:

• Sincast successfully projects single cells into correct biological niches within bulk reference atlases.
• Demonstrated the effectiveness of the imputation approach for scRNA-seq querying.
• Showcased Sincast's ability to avoid batch effect correction issues.

Conclusions:

• Sincast is an efficient and powerful tool for single-cell profiling.
• Facilitates downstream analysis of scRNA-seq data by leveraging bulk reference atlases.
• Enables accurate cell identity prediction and discovery of new cell states.