TarPmiR

TarPmiR predicts microRNA (miRNA) target sites on messenger RNA (mRNA) using CLASH-derived data and machine learning to improve identification of miRNA-mRNA interactions.

Key Features:

• CLASH data analysis: Leverages CLASH (cross-linking ligation and sequencing of hybrids) experimental data to derive direct miRNA-mRNA interaction information.
• Machine learning approaches: Applies four distinct machine learning approaches to analyze CLASH-derived data.
• Seven novel features: Identifies seven novel features from CLASH data that enhance prediction accuracy of miRNA-mRNA interactions.
• Feature integration: Integrates the seven novel features with traditional target-site features used by existing algorithms to build a combined predictive model.
• Validation on non-CLASH datasets: Validated on two human and one mouse non-CLASH datasets, predicting over 74.2% of true miRNA target sites across each dataset.
• Comparative performance: Outperforms three established prediction approaches in comparative analyses, exhibiting improved recall and precision.

Scientific Applications:

• Gene regulation studies: Enables more accurate identification of miRNA-mRNA interactions for investigations of post-transcriptional gene regulation.
• Post-transcriptional network analysis: Supports mapping and analysis of miRNA-mediated regulatory networks.
• Developmental biology research: Facilitates study of miRNA roles in developmental processes via improved target-site prediction.
• Disease research: Assists in identifying miRNA-target interactions relevant to disease mechanisms.
• Therapeutic target discovery: Aids in prioritizing miRNA-mRNA interactions for potential therapeutic intervention studies.

Methodology:

Analyzed CLASH data; applied four distinct machine learning approaches to identify seven novel features; integrated those features with traditional target-site features; validated predictions on two human and one mouse non-CLASH datasets and performed comparative analyses against three established prediction methods.