CorEx

CorEx identifies groups of differentially expressed genes in tumor RNA-sequencing datasets to extract correlated gene expression factors for precision oncology analyses.

Key Features:

• Machine Learning Algorithm: The CorEx algorithm identifies and models groups of genes exhibiting differential expression and extracts latent factors from tumor RNA-seq gene expression data.
• CorExplorer component: CorExplorer links generated factors to external datasets and supports factor-level analysis and interpretation.
• Multi-Tumor Analysis: Models are trained on RNA-seq gene expression data from ovarian, lung, melanoma, and colorectal tumors.
• Integration of External Datasets: Integrates survival data, protein-protein interactions, Gene Ontology (GO) and KEGG pathway enrichment results to contextualize factors.
• Heatmap Visualization: Generates heatmaps to visualize gene expression patterns across identified factors.
• Factor Graph Visualization: Produces factor graph visualizations showing relationships among gene expression factors and their associations with external datasets.

Scientific Applications:

• Precision Oncology: Facilitates analysis of differential gene expression in tumors to inform precision oncology studies.
• Tumor Biology and Target Discovery: Enables identification of biologically coherent gene modules, pathway associations, prognostic signals, and candidate therapeutic targets.

Methodology:

Training the CorEx algorithm on RNA-seq gene expression data from ovarian, lung, melanoma, and colorectal tumors, followed by analysis of resulting factors with CorExplorer and integration of survival data, protein-protein interactions, GO and KEGG enrichment results and heatmaps.