Correlation-guided Network Integration (CoNI), an R package for integrating numerical omics data that allows...

Area of Science:

• Bioinformatics
• Computational Biology
• Data Science

Background:

• The rapid expansion of complex biological data necessitates advanced tools for integration and analysis.
• Extracting meaningful insights from omics datasets presents a significant challenge.

Purpose of the Study:

• To introduce CoNI, a practical R package designed for unsupervised integration of numerical omics datasets.
• To provide a flexible tool for identifying confounding variables and building integrated biological networks.

Main Methods:

• CoNI employs partial correlations to detect potential confounding variables in paired omics data.
• The package constructs integrated networks, representing them as weighted undirected graphs, bipartite graphs, or hypergraphs.
• This network-based approach facilitates further biological data analysis.

Main Results:

• CoNI enables the unsupervised integration of multiple omics datasets.
• The generated network structures can be used to identify biologically significant candidate variables.
• Network comparisons across different conditions are possible.

Conclusions:

• CoNI offers a robust and flexible solution for omics data integration and analysis.
• The package supports the identification of key biological insights through network representation.
• CoNI enhances the ability to explore complex biological relationships within integrated omics data.