Gene trajectory inference for single-cell data by optimal transport metrics.

Area of Science:

• Computational Biology
• Genomics
• Systems Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) is crucial for studying cell state transitions and gene dynamics.
• Current cell trajectory inference methods using cell pseudotime can be confounded by concurrent gene processes and technical noise.
• Accurate inference of gene dynamics is essential for understanding complex biological processes.

Purpose of the Study:

• To introduce GeneTrajectory, a novel computational approach for inferring gene dynamics.
• To overcome the limitations of cell-centric trajectory inference methods.
• To enable the discovery of gene programs driving biological processes.

Main Methods:

• GeneTrajectory calculates optimal transport distances between gene distributions across a cell-cell graph.
• It extracts gene programs and defines their pseudotemporal order.
• The approach focuses on gene trajectories rather than cell trajectories.

Main Results:

• GeneTrajectory accurately captures progressive gene dynamics during myeloid lineage maturation.
• It successfully deconvolves key gene programs in mouse hair follicle differentiation, outperforming cell trajectory methods.
• The method demonstrates robust identification of dynamic gene expression patterns.

Conclusions:

• GeneTrajectory provides a powerful new tool for analyzing gene dynamics in scRNA-seq data.
• It enhances the understanding of complex biological processes by resolving gene programs.
• This approach facilitates the discovery of regulatory gene networks underlying cell differentiation and development.