scMUG: deep clustering analysis of single-cell RNA-seq data on multiple gene functional modules.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Single-cell RNA sequencing (scRNA-seq) provides high-resolution gene expression data, revealing cellular heterogeneity.
• Analyzing scRNA-seq data is challenging due to sparsity and high dimensionality.
• Bioinformatics is crucial for analyzing large biological datasets like scRNA-seq.

Purpose of the Study:

• Introduce the scMUG computational pipeline to improve scRNA-seq clustering analysis.
• Integrate gene functional module information into scRNA-seq data analysis.
• Address the challenges of sparsity and high dimensionality in scRNA-seq data.

Main Methods:

• Developed the scMUG pipeline encompassing data preprocessing, cell representation, similarity matrix construction, and clustering.
• Introduced a novel similarity measure combining local density and global distribution in latent space.
• Integrated gene functional associations into the clustering process.

Main Results:

• The scMUG pipeline demonstrated enhanced clustering performance on nine human scRNA-seq datasets.
• Integration of gene functional information provided deeper insights into cellular heterogeneity.
• scMUG achieved comparable or superior results compared to existing state-of-the-art methods.

Conclusions:

• The scMUG pipeline effectively leverages gene functional modules for robust scRNA-seq clustering.
• This approach offers a novel way to understand functional relationships in cellular heterogeneity.
• The pipeline provides a valuable tool for scRNA-seq data analysis with publicly available code.