robin2: accelerating single-cell data clustering evaluation.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Single-cell RNA sequencing (scRNA-seq) technologies are rapidly advancing, necessitating improved methods for evaluating data clustering.
• Existing clustering evaluation methods may lack scalability and efficiency for high-dimensional scRNA-seq datasets.

Purpose of the Study:

• To develop an optimized R package, robin2, for robust and scalable evaluation of clustering in scRNA-seq data.
• To enhance computational efficiency and network analysis capabilities for large-scale biological datasets.

Main Methods:

• Development of robin2, an optimized version of the R package robin.
• Implementation of enhanced computational efficiency and support for high-dimensional datasets.
• Integration with R's base functionalities for network analysis and clustering stability validation.

Main Results:

• robin2 demonstrates improved functionality for clustering stability validation and systematic evaluation of community detection algorithms.
• Application to Tabula Muris and PBMC datasets successfully identified biologically meaningful cell subpopulations with high statistical significance.
• The new version achieves a 9-fold reduction in computational time on large-scale datasets through parallel processing.

Conclusions:

• robin2 provides a robust, efficient, and scalable solution for evaluating clustering in scRNA-seq data.
• The package facilitates the identification of significant cell subpopulations and improves network analysis.
• robin2 is freely available on CRAN with comprehensive documentation and a detailed analysis vignette.