Interpretable trajectory inference with single-cell linear adaptive negative-binomial expression (scLANE) testing.

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Single-cell RNA sequencing (scRNA-seq) enables gene expression dynamics analysis.
• Trajectory inference methods identify cell differentiation paths.
• Existing differential expression methods for trajectories are often nonlinear and difficult to interpret.

Purpose of the Study:

• To develop a novel, interpretable statistical method for trajectory differential expression analysis in scRNA-seq data.
• To address the interpretability challenges posed by nonlinear models in current trajectory analysis.

Main Methods:

• Proposed single-cell linear adaptive negative-binomial expression (scLANE) testing.
• Utilized an interpretable generalized linear model framework.
• Incorporated basis splines for handling nonlinear gene expression dynamics.
• Extended methods for complex experimental designs using estimating equations and mixed models.

Main Results:

• Validated scLANE's accuracy through various simulation scenarios.
• Demonstrated scLANE's ability to extract novel biological information from diverse scRNA-seq datasets.
• Showcased scLANE's utility with both pseudotime and RNA velocity estimation methods.

Conclusions:

• scLANE provides a robust and interpretable framework for trajectory differential expression analysis.
• The method enhances biological discovery from scRNA-seq data, particularly in complex differentiation processes.
• scLANE is available as an R package and a web server for broader accessibility.