Identifying Single-Cell Expression Quantitative Trait Loci Using a Bootstrap Penalized Hurdle Model

Area of Science:

• Genomics
• Computational Biology
• Statistical Genetics

Background:

• Expression quantitative trait loci (eQTL) analysis connects genetic variants to gene expression, aiding the study of gene regulation.
• Single-cell RNA sequencing (scRNA-seq) enables cell-type-specific eQTL detection, but data sparsity and heterogeneity pose challenges.
• Existing methods struggle with the complexities of scRNA-seq data for eQTL mapping.

Purpose of the Study:

• To introduce a novel statistical framework, BPHurdle, tailored for scRNA-seq data.
• To address the challenges of sparsity and heterogeneity in single-cell eQTL analysis.
• To improve the accuracy and robustness of identifying cell-type-specific regulatory variants.

Main Methods:

• Developed the Bootstrap Penalized Hurdle regression model (BPHurdle).
• Utilized a hurdle framework with logistic and Poisson components to model excess zeros and positive expression levels.
• Applied the model to both simulated and real scRNA-seq datasets.

Main Results:

• BPHurdle demonstrated high accuracy and robustness in simulations for identifying regulatory variants.
• Successfully identified reliable cell-type-specific eQTLs in a real scRNA-seq dataset case study.
• Validated the model's effectiveness on differentially expressed genes.

Conclusions:

• BPHurdle provides an advanced and flexible approach for single-cell eQTL mapping.
• Offers deeper insights into genetic regulation of gene expression at cellular resolution.
• Facilitates more precise understanding of genetic influences on cellular function.