RPmirDIP

RPmirDIP predicts miRNA–gene interactions using a cascaded semi-supervised machine learning framework that refines existing mirDIP scores through reciprocal perspective analysis.

Key Features:

• Reciprocal Perspective Method: Applies local decision thresholds derived from dual complementary perspectives for each miRNA–gene pair to improve target prediction accuracy over global threshold approaches.
• Large-Scale Interaction Refinement: Reassesses approximately 6 million unique miRNA–gene pairs from mirDIP to generate an expanded and refined interaction dataset.

Scientific Applications:

• miRNA-Mediated Gene Regulation Analysis: Supports investigation of regulatory networks and identification of disease-associated miRNA–gene interactions.

Methodology:

RPmirDIP augments mirDIP prediction scores using a cascaded semi-supervised learning approach that evaluates each miRNA–gene pair from reciprocal perspectives and applies pair-specific local thresholds to refine interaction probabilities.