birta

birta infers transcription factor (TF) and microRNA (miRNA) activities from combined mRNA and miRNA expression data by integrating TF-target and miRNA-target network annotations to produce condition-specific regulatory activity inferences.

Key Features:

• Bayesian Network Modeling: Employs a Bayesian network framework to model probabilistic regulatory relationships among TFs, miRNAs, and target genes.
• Markov-Chain-Monte-Carlo Sampling: Uses Markov-Chain-Monte-Carlo (MCMC) sampling to explore posterior activity states within the Bayesian network.
• Condition-Specific Inference: Infers regulator activities tailored to experimental conditions and distinguishes positive switches (inactive to active) and negative switches (active to inactive).
• Performance Validation: Validated by extensive simulations demonstrating superior prediction performance relative to other approaches.
• Application Examples: Applied to Escherichia coli under aerobic and anaerobic growth conditions and to human pancreas and ovarian cancer expression datasets.

Scientific Applications:

• Elucidating regulatory mechanisms: Provides insights into how TFs and miRNAs govern condition-specific gene expression programs.
• Systems biology, genomics, and personalized medicine: Supports studies in systems biology, genomics, and personalized medicine by identifying condition-specific regulatory activity relevant to disease mechanisms and therapeutic targets.

Methodology:

Integrates TF-target and miRNA-target annotations with mRNA and miRNA expression data to construct a Bayesian network and applies MCMC sampling to explore activity states for prediction of regulatory switches.