Challenges and progress in RNA velocity: Comparative analysis across multiple biological contexts.

Area of Science:

• Genomics
• Computational Biology
• Molecular Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) enables transcriptomic profiling of individual cells.
• RNA velocity is a computational method predicting future cell states from mRNA dynamics.
• Understanding cell differentiation and disease requires accurate trajectory inference.

Purpose of the Study:

• To benchmark and compare the performance of five RNA velocity methods.
• To evaluate methods based on local consistency, agreement, driver gene identification, and robustness.
• To provide guidance for selecting appropriate RNA velocity tools in biological research.

Main Methods:

• Development of a comparative analysis pipeline for RNA velocity methods.
• Application of the pipeline to three distinct scRNA-seq datasets.
• Performance evaluation using metrics including local consistency, method agreement, driver gene identification, and robustness to sequencing depth.

Main Results:

• Significant performance variations were observed among the five RNA velocity methods.
• Method accuracy was dataset-dependent, highlighting the need for careful selection.
• Robustness to sequencing depth differed across methods, impacting reliability.

Conclusions:

• No single RNA velocity method excels in all scenarios.
• The benchmark provides essential insights into the strengths and limitations of current RNA velocity tools.
• This resource aids researchers in choosing optimal methods for their specific scRNA-seq studies.