simplifyEnrichment

simplifyEnrichment clusters and visualizes functional enrichment results to reduce redundancy and improve interpretation of Gene Ontology (GO) and gene set enrichment analyses.

Key Features:

• Implementation: Implemented as an R/Bioconductor package.
• Binary cut clustering: Implements the "binary cut" algorithm to partition terms into mutually exclusive clusters with high internal similarity.
• Semantic similarity-based clustering: Clusters GO terms using semantic similarity matrices.
• Alternative similarity comparison: Compares semantic similarity matrices to gene overlap-based similarity matrices.
• Partitioning assessment: Supports assessment of various partitioning methods within the binary cut framework.
• Benchmarking: Benchmarked on simulated and real-world datasets to evaluate clustering coherence.
• Visualization: Produces word clouds to summarize and visualize terms within each cluster.
• Redundancy reduction: Reduces redundant enrichment terms to simplify result interpretation.
• Input scope: Operates on results from functional enrichment analysis and gene set enrichment analysis.

Scientific Applications:

• Enrichment result summarization: Summarizing and interpreting GO and gene set enrichment analysis outputs.
• Redundancy minimization: Reducing redundancy in lists of significant functional terms to aid biological interpretation.
• Method comparison: Comparing similarity measurements (semantic similarity vs gene overlap) and partitioning strategies to optimize analysis workflows.
• Clustering evaluation: Benchmarking clustering performance on simulated and real datasets to assess cluster coherence.
• Visualization for interpretation: Visualizing functional term clusters with word clouds to facilitate biological insight.

Methodology:

Compute semantic similarity matrices (and optionally gene overlap matrices), apply the binary cut clustering algorithm to GO terms, compare different similarity measurements and partitioning methods, benchmark on simulated and real datasets, and visualize clusters using word clouds; implemented in R/Bioconductor.