DiffCorr: an R package to analyze and visualize differential correlations in biological networks

Area of Science:

• Bioinformatics
• Systems Biology
• Genomics

Background:

• Large-scale omics data (e.g., microarrays) are crucial for understanding cellular regulatory networks and the molecular basis of diseases.
• Correlation analysis, including hierarchical clustering, is a common method for analyzing omics data.
• Changes in molecule correlations, not just mean levels, between experimental conditions are vital for biological insights.

Purpose of the Study:

• To develop an efficient and unbiased tool for identifying differential correlation patterns in omics data.
• To provide a method for analyzing changes in correlation networks between two experimental conditions.

Main Methods:

• Developed the DiffCorr package, a method for identifying differential correlation patterns.
• Utilizes Pearson's correlation coefficient as the association measure.
• Calculates correlation matrices, identifies principal component-based 'eigen-molecules', and applies Fisher's z-test for differential correlation analysis.

Main Results:

• Demonstrated DiffCorr's utility in identifying biologically relevant, differentially correlated molecules in transcriptome and metabolome data.
• Showcased the ability to explore differential correlations across different post-genomics data types (transcriptomics, metabolomics).

Conclusions:

• DiffCorr offers a simple and effective approach for calculating differential correlations in omics data.
• The package serves as a valuable first step towards inferring causal relationships and identifying potential biomarker candidates.