runibic: a Bioconductor package for parallel row-based biclustering of gene expression data.

Area of Science:

• Bioinformatics
• Computational Biology
• Data Mining

Background:

• Biclustering is a simultaneous row and column clustering technique for matrices.
• UniBic is a highly accurate biclustering algorithm.
• Existing implementations may lack efficiency and modern integration.

Purpose of the Study:

• Introduce the runibic R package for biclustering.
• Provide a parallelized and optimized implementation of the UniBic algorithm.
• Facilitate integration with modern bioinformatics data structures.

Main Methods:

• Reimplementation and parallelization of the UniBic algorithm in R.
• Development of the runibic package for Bioconductor.
• Integration with Bioconductor data structures like SummarizedExperiment and ExpressionSet.

Main Results:

• runibic offers a parallelized UniBic version significantly faster than the sequential algorithm.
• Improved memory management enhances computational efficiency.
• The package provides modularity for developing new biclustering methods.

Conclusions:

• The runibic package provides an efficient and user-friendly tool for biclustering.
• It enables faster and more scalable analysis of biological data.
• runibic supports integration with the broader Bioconductor ecosystem.