Integrating large-scale single-cell RNA sequencing in central nervous system disease using self-supervised...

Area of Science:

• Neuroscience
• Computational Biology
• Genomics

Background:

• The central nervous system (CNS) contains diverse cell types crucial for its function.
• Single-cell RNA sequencing (scRNA-seq) offers insights into brain cell atlases.
• Integrating large-scale CNS scRNA-seq data is challenging due to cell heterogeneity.

Purpose of the Study:

• To develop a novel method for integrating large-scale CNS scRNA-seq datasets.
• To reveal heterogeneous relationships within CNS cell types and subtypes.
• To create a reference atlas for cell type and subtype annotation in neural tissues.

Main Methods:

• Introduced scCM, a self-supervised contrastive learning method.
• scCM compares gene expression variations to group similar cells and separate dissimilar ones.
• Evaluated scCM on 20 CNS datasets across 4 species and 10 CNS diseases.

Main Results:

• scCM effectively integrates heterogeneous CNS scRNA-seq data.
• The method accurately annotates cell types and subtypes in neural tissues.
• Rich spatial information of cell states was obtained, aiding CNS disease research.

Conclusions:

• scCM is a robust and promising method for large-scale CNS scRNA-seq data integration.
• This approach enhances understanding of cellular and molecular mechanisms in CNS functions and diseases.
• Facilitates the creation of comprehensive CNS cell atlases.