Human Gene Correlation Analysis (HGCA)

Human Gene Correlation Analysis (HGCA) analyzes global gene coexpression across 35,000 GTEx RNA-Seq samples to identify coexpressed gene subclades and infer functional and regulatory relationships.

Key Features:

• Dataset: Uses 35,000 RNA-Seq samples sourced from the GTEx Project encompassing multiple healthy human tissues.
• Hierarchical clustering: Applies hierarchical clustering to gene expression profiles to define coexpression structure and subclades.
• Gene coexpression network (GCN): Constructs a gene coexpression network from calculated gene similarity based on expression values.
• Subclade extraction: Presents subclades of coexpressed genes for any given gene of interest.
• Enrichment analyses: Integrates enrichment for gene ontologies, biological pathways, protein families, and diseases to interpret clusters.
• Transcription factor enrichment: Identifies enriched transcription factors that may drive observed coexpression patterns.
• Benchmarking: Evaluates performance against other coexpression webtools using STRING analysis.

Scientific Applications:

• Functional association inference: Hypothesizes functional relationships between genes based on coexpression.
• Gene function prediction: Prioritizes candidate genes for functional annotation by clustering with annotated genes.
• Regulatory network investigation: Infers potential regulatory mechanisms and transcription factor drivers of coexpressed modules.
• Experimental hypothesis generation: Generates testable hypotheses about gene partnerships and shared biological processes for laboratory validation.
• Tissue-specific expression analysis: Supports exploration of ubiquitous versus tissue-specific gene expression patterns across GTEx tissues.

Methodology:

Uses 35,000 GTEx RNA-Seq samples; preprocesses transcriptomic data; calculates gene similarity from expression values; applies hierarchical clustering; constructs a gene coexpression network (GCN); evaluates clusters via biological term enrichment analysis; and benchmarks results using STRING analysis.