RANKS

RANKS implements node label ranking and classification in biomolecular networks using kernelized score functions for tasks such as automated protein function prediction, gene disease prioritization, and drug repositioning.

Key Features:

• Node label ranking and classification: RANKS performs ranking and classification of nodes within biological and biomolecular networks.
• Kernelized score functions: The package implements kernelized score functions as the core scoring mechanism.
• Semi-supervised algorithmic framework: RANKS applies a semi-supervised framework that integrates both local and global learning strategies.
• Local and global learning strategies: Local and global learning components are combined to enable comprehensive network analysis.
• Baseline network-based methods: Implementations of label propagation and random walk algorithms are included for comparative assessment.
• Implementation languages: Core functionalities are implemented in R with computationally intensive components implemented in C.

Scientific Applications:

• Automated protein function prediction: RANKS can be applied to predict protein functions by ranking proteins within interaction networks.
• Gene disease prioritization: The method supports prioritization of candidate genes related to disease phenotypes via network-based ranking.
• Drug repositioning: RANKS can be used to identify repositioning opportunities by ranking drugs or targets in biological networks.
• Comparative method assessment: Inclusion of label propagation and random walk baselines facilitates benchmarking against established network-based techniques.

Methodology:

Implements kernelized score functions within a semi-supervised algorithmic framework that combines local and global learning strategies and includes label propagation and random walk algorithms, with R implementation and computationally intensive parts in C.