mirTime

mirTime identifies condition-specific microRNA (miRNA) targets from time-series transcript data by integrating sequence features and temporal expression profiles to detect dynamic regulatory interactions.

Key Features:

• Integration of Sequence and Expression Data: combines sequence features and time-series expression profiles to predict miRNA targets within specific experimental contexts.
• Gaussian Process Regression: applies Gaussian process regression to estimate expression values at unobserved or unpaired time points and model temporal profiles.
• Condition-Specific Target Modules: detects target modules specific to experimental conditions and cell types by leveraging datasets from multiple conditions.
• Correlation-Based Inference: exploits negative correlation between miRNA and target expression across time-series to prioritize candidate targets.
• Comprehensive Evaluation Criteria: assesses performance using correlation-based, target gene-based, and network-based evaluation criteria.

Scientific Applications:

• Dynamic regulatory network analysis: elucidates miRNA–target regulatory networks across time points and experimental settings.
• Disease and developmental studies: provides insights into disease mechanisms and developmental processes by identifying condition-specific miRNA targets.
• Environmental and perturbation responses: characterizes temporal miRNA-mediated responses to environmental changes or experimental perturbations.

Methodology:

Initial target candidate selection is based on sequence features and binding energy, and target prediction prioritizes negatively correlated miRNA–target pairs using time-series expression analysis enhanced by Gaussian process regression.