parmigene

parmigene reconstructs gene regulatory networks by estimating mutual information from gene expression data using a k-nearest neighbor entropy estimator implemented in R and parallelized for large-scale analysis.

Key Features:

• k-nearest neighbor mutual information estimator: Uses k-nearest neighbor distances to compute entropy-based mutual information estimates with reduced bias.
• Parallel computing: Parallelizes mutual information computations to scale to thousands of genes and reduce computational time.
• R package implementation: Implemented as an R package (parmigene) for integration with R-based analysis workflows.
• Empirical validation: Demonstrated accuracy and computational efficiency on in silico (simulated) and real-world gene expression datasets.

Scientific Applications:

• Gene regulatory network reconstruction: Infers transcriptional networks from gene expression data using mutual information.
• Large-scale transcriptional analysis: Enables analysis of genome-scale expression datasets by reducing computational costs of mutual information estimation.
• Study of biological processes and disease mechanisms: Supports exploration of complex biological processes and disease-related transcriptional regulation.

Methodology:

The method estimates mutual information via entropy calculations derived from k-nearest neighbor distances and performs these computations within a parallel computing framework.