CorrAdjust

CorrAdjust corrects for confounding variables when computing correlations between transcriptomic features by selecting principal components to residualize from expression data, enhancing biological relevance and interpretability.

Key Features:

• Confounder Correction: Identifies and corrects hidden confounders in RNA-RNA correlation computations by residualizing selected principal components from expression data.
• Enrichment-Based Metric: Uses a novel enrichment-based metric that maximizes the enrichment of reference pairs among highly correlated RNA-RNA pairs, specifically designed to evaluate correlation data and provide RNA-level interpretability.
• Cross-Dataset Validation: Evaluated on 25,063 human RNA-seq datasets from The Cancer Genome Atlas, the Genotype-Tissue Expression project, and the Geuvadis collection.
• Small RNA and mRNA Integration: Excels at integrating small RNA and mRNA sequencing data, significantly enhancing the enrichment of experimentally validated miRNA targets among negatively correlated miRNA-mRNA pairs.

Scientific Applications:

• Transcriptomic Correlation Analysis: Computes biologically relevant correlations between transcriptomic features by eliminating hidden confounders from RNA-seq data.
• miRNA-mRNA Interaction Studies: Enhances detection of experimentally validated miRNA targets through improved identification of negatively correlated miRNA-mRNA pairs.
• Multi-Dataset Transcriptomics: Applies across diverse human RNA-seq datasets including cancer, tissue-specific, and population-scale transcriptomic studies.

Methodology:

CorrAdjust selects a subset of principal components to residualize from expression data by maximizing the enrichment of reference pairs among highly correlated RNA-RNA pairs.