scENT for Revealing Gene Clusters From Single-Cell RNA-Seq Data.

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Single-cell RNA sequencing (scRNA-seq) allows high-resolution transcriptomic analysis of individual cells.
• Understanding gene-disease relationships is crucial for biological research.
• Existing methods primarily focus on cell-level clustering, lacking gene-level insights.

Purpose of the Study:

• To introduce scENT (single cell gENe clusTer), a novel deep learning framework for identifying biologically significant gene clusters from scRNA-seq data.
• To address challenges in high-dimensional scRNA-seq data, including sparsity and dropout.
• To facilitate the discovery of novel functional gene clusters and their associated functions.

Main Methods:

• Clustering scRNA-seq data into optimal groups.
• Performing gene set enrichment analysis to identify over-represented gene classes.
• Integrating perturbation into the deep learning clustering process for robustness.

Main Results:

• scENT demonstrated superior performance compared to existing methods on simulation data.
• Application to Alzheimer's disease and brain metastasis datasets revealed novel functional gene clusters.
• The identified clusters provided insights into prospective disease mechanisms.

Conclusions:

• scENT is an effective deep learning framework for identifying significant gene clusters in scRNA-seq data.
• The method enhances the biological interpretation of scRNA-seq data.
• scENT aids in understanding disease mechanisms and discovering potential therapeutic targets.