Bulk WGCNA analysis with omicverse

WGCNA (Weighted Gene Co-expression Network Analysis) identifies gene co-expression modules by constructing adjacency and topological overlap matrices from expression data, then detecting modules through dendrogram clustering. This reveals groups of genes with coordinated expression patterns across your samples.

WGCNA requires filtering expression data using MAD (median absolute deviation), selecting variable genes, and ensuring data quality. These preprocessing steps remove noise and focus analysis on biologically informative genes before network construction.

Yes, WGCNA enables module–trait analysis by calculating eigengenes (representative expression profiles for each module) and correlating them with sample metadata. This reveals which gene modules associate with clinical phenotypes or disease progression.

Hub gene extraction ranks genes within a module by connectivity and correlation strength, then sub-network plotting visualizes the top hub genes and their relationships. This identifies the most central genes driving module function.

Yes, WGCNA is designed for high-dimensional bulk RNA-seq data with many genes and samples. It efficiently handles large datasets and is implemented here via omicverse for streamlined analysis on complex transcriptomic studies.

Soft-thresholding applies weighted power transformation to correlation values, preserving continuous information. This constructs the adjacency matrix, which feeds into topological overlap matrix (TOM) calculation to improve module detection reliability over hard thresholding.