SiDCo

SiDCo computes distance and signed distance correlations to quantify linear and non-linear associations among variables for constructing data-driven biological networks from omics datasets.

Key Features:

• Distance Correlation Calculation: Computes distance correlation to capture both linear and non-linear dependencies between variables.
• Signed Distance Correlation: Integrates the sign from Pearson's correlation with distance correlation values to provide directional information on associations.
• Flexible Data Handling: Supports correlations between vectors of different lengths and offers one-to-one or one-to-all correlation modes.
• Partial Distance Correlation: Implements partial distance correlation using a Gaussian Graphical model approach adapted to distance covariance to control for confounders.

Scientific Applications:

• Data-driven network development: Construction of biological association networks to study interactions among metabolites and other molecular features.
• Omics association analysis: Analysis of linear and non-linear dependencies in genomics, proteomics, metabolomics, and lipidomics datasets.

Methodology:

Computes distance correlation; derives signed distance correlation by combining distance correlation magnitudes with the sign from Pearson's correlation; supports unequal-length vector correlations and one-to-one/one-to-all modes; computes partial distance correlations via a Gaussian Graphical model adapted to distance covariance.