EnGRaiN

EnGRaiN reconstructs genome-scale gene regulatory networks from gene expression data using supervised ensemble learning.

Key Features:

• Supervised Ensemble Learning: Employs supervision with labeled positive and negative gene pair interactions during training to improve prediction accuracy.
• Integration of Multiple Methods: Combines predictions from various network inference methods into a unified ensemble prediction.
• Performance Metrics (ROC and PR): Demonstrates improved receiver operating characteristic (ROC) and precision-recall (PR) characteristics on simulated datasets and curated Arabidopsis thaliana microarray data.
• Biological Insight Generation: Enables mining of reconstructed networks to uncover complex gene regulatory interactions and pathways.

Scientific Applications:

• Gene Expression Analysis: Infers regulatory relationships from gene expression datasets, including microarray data.
• Biological Interaction Elucidation: Supports discovery of novel gene regulatory interactions and pathways.
• Comparative Genomics: Facilitates comparison of gene regulatory networks across species or conditions for evolutionary and functional genomics studies.

Methodology:

Data preparation using simulated datasets and curated Arabidopsis thaliana microarray data; supervised training with labeled positive and negative gene pair interactions; ensemble network construction by combining predictions from multiple network inference methods; validation using ROC and PR metrics.