Integrating single-cell datasets with ambiguous batch information by incorporating molecular network features.

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Single-cell RNA sequencing (scRNA-seq) enables large-scale cell atlas projects.
• Integrating diverse scRNA-seq datasets is challenging due to batch effects and data heterogeneity.

Purpose of the Study:

• To develop a robust and scalable method for integrating scRNA-seq datasets.
• To address challenges in combining data from varied tissue sources and developmental stages.

Main Methods:

• SCORE, a network-based integration methodology.
• Incorporation of curated molecular network features.
• Inference of cellular states for unified data integration.

Main Results:

• SCORE effectively integrates scRNA-seq datasets regardless of batch information.
• Demonstrated superior accuracy, robustness, scalability, and data integration compared to existing methods.
• Validated performance on real-world single-cell datasets.

Conclusions:

• SCORE provides a unified workflow for scRNA-seq data integration.
• The method enhances the ability to build comprehensive cell atlases.
• SCORE offers a significant advancement in handling batch effects in scRNA-seq analysis.