RANKS: a flexible tool for node label ranking and classification in biological networks

Area of Science:

• Bioinformatics
• Computational Biology
• Network Analysis

Background:

• Bioinformatics tasks often involve ranking network nodes based on specific properties.
• Existing methods may lack flexibility or efficiency for complex biomolecular network analysis.

Purpose of the Study:

• Introduce RANKS, a versatile software package for bioinformatics network analysis.
• Provide an efficient and user-friendly implementation of semi-supervised learning for node ranking.

Main Methods:

• Utilizes kernelized score functions with integrated local and global learning strategies.
• Implements semi-supervised learning for analyzing biomolecular networks.
• Includes baseline network-based methods like label propagation and random walks for comparison.

Main Results:

• RANKS offers a flexible framework applicable to diverse bioinformatics problems.
• The package provides efficient computational performance through C implementation for intensive functions.
• Facilitates comparative analysis by including standard network-based algorithms.

Conclusions:

• RANKS is a valuable tool for automated protein function prediction, gene disease prioritization, and drug repositioning.
• The software's design supports both local and global learning strategies in network analysis.
• Availability on CRAN and supplementary data enhance its utility for the research community.