Statistical methods for analysis of single-cell RNA-sequencing data.

Area of Science:

• Genomics
• Computational Biology
• Statistical Genetics

Background:

• Single-cell RNA sequencing (scRNA-seq) provides high-resolution gene expression data.
• Analyzing scRNA-seq data is complex due to biological variability and technical artifacts like dropout events.
• Existing methods often struggle to account for the molecular capture process inherent in scRNA-seq.

Purpose of the Study:

• To develop a novel statistical approach for analyzing single-cell RNA sequencing Unique Molecular Identifier (UMI) counts.
• To address challenges in scRNA-seq data analysis, including dropout events and molecular capture variability.
• To enable robust downstream analyses such as differential expression and cell type detection.

Main Methods:

• A new statistical model is proposed for fitting and analyzing observed scRNA-seq UMI count data.
• The method incorporates the molecular capture process and biological factors leading to dropout events.
• External spike-in data can be utilized to improve model accuracy.
• Cellular auxiliary information (clusters, cell cycle) is used to mitigate unwanted variation.

Main Results:

• The developed approach accurately models scRNA-seq UMI data, including zero-inflated counts.
• Differential expression analysis is performed with adjustments for cell capture rates.
• The method facilitates differential zero inflation analysis, influential gene classification, and marker gene selection.
• Improved cell type detection and estimation of capture rates are achieved.

Conclusions:

• The novel statistical approach provides a robust framework for comprehensive analysis of scRNA-seq UMI count data.
• By accounting for dropout events and capture rates, the method enhances the reliability of gene expression studies.
• This approach offers improved tools for downstream analyses, leading to more accurate biological insights from scRNA-seq experiments.