JAMI

JAMI computes conditional mutual information (CMI) from genome-wide miRNA and gene expression data to support identification and analysis of competing endogenous RNA (ceRNA) networks.

Key Features:

• Conditional Mutual Information (CMI) computation: Computes CMI values between miRNA and gene expression variables for ceRNA analysis.
• Non-parametric estimator: Uses a non-parametric estimator to calculate CMI values from expression data.
• High performance: Reports an approximately 70-fold speedup compared to previous implementations.
• Multi-threading: Supports multi-threaded execution to leverage multi-core architectures for faster computation.
• Java 8 implementation: Implemented in Java 8 as the computational runtime environment.
• Scalability: Designed to handle large-scale, genome-wide miRNA and gene expression datasets.

Scientific Applications:

• ceRNA network inference: Identification and analysis of competing endogenous RNA (ceRNA) interactions from paired miRNA and gene expression data.
• Cross-class RNA analysis: Analysis of ceRNA relationships involving both protein-coding and non-coding RNAs.
• Genome-wide expression studies: Application to large-scale genome-wide miRNA and gene expression datasets for regulatory network discovery.
• Investigation of regulatory mechanisms in health and disease: Support for studies probing RNA-mediated gene regulation and its roles in healthy and diseased states.

Methodology:

Computes conditional mutual information using a non-parametric estimator implemented in Java 8 with multi-threading, yielding an approximately 70-fold speedup over prior implementations.