bluster

bluster wraps common clustering algorithms in an easily extended S4 framework for comparing and evaluating clustering results.

Key Features:

• Hierarchical clustering backend: Implements hierarchical clustering as one of the supported algorithm backends.
• K-means clustering backend: Implements k-means clustering as one of the supported algorithm backends.
• Graph-based clustering backend: Implements graph-based clustering as one of the supported algorithm backends.
• Clustering comparison utilities: Provides utilities to compare and evaluate clustering results.
• Extensible S4 framework: Enables easy extension through an S4 object-oriented framework.

Scientific Applications:

• Clustering algorithm benchmarking: Enables systematic comparison and evaluation of different clustering approaches on biological data.
• Flexible clustering pipelines: Supports integration of multiple clustering strategies within Bioconductor workflows for single-cell and other high-dimensional biological data.

Methodology:

The package implements an S4 framework with pluggable backends for hierarchical clustering, k-means clustering, and graph-based clustering, supplemented by utilities for comparing and evaluating clustering results.