Mirsynergy

Mirsynergy identifies microRNA regulatory modules (MiRMs) by clustering co-occurring miRNA targets and optimizing a synergy score derived from miRNA–mRNA and gene–gene interactions to reveal functionally coherent regulatory modules.

Key Features:

• Deterministic Overlapping Clustering Algorithm: Employs a deterministic overlapping clustering algorithm adapted from a recently developed framework, avoiding stochastic methods and not requiring a predetermined number of regulatory modules.
• Two-Stage Operational Framework: Forms MiRMs from the co-occurrence of miRNA targets and subsequently expands each MiRM using a greedy algorithm that includes or excludes mRNAs to maximize a synergy score computed from miRNA–mRNA and gene–gene interactions.
• Synergy Score Optimization: Optimizes a synergy score as the primary criterion to evaluate functional coherence and interaction strength within identified MiRMs.

Scientific Applications:

• Functional Enrichment Analysis: Mirsynergy-MiRMs show significantly higher functional enrichment compared with internal controls and existing methods.
• Expression Correlation Coherence: Identified MiRMs exhibit coherent anti-correlation between miRNA and mRNA expression patterns, supporting biological relevance.
• Prognostic Utility in Cancer Research: Through Kaplan-Meier survival analysis, Mirsynergy has proposed several prognostically promising MiRMs relevant to cancer outcomes.

Methodology:

Mirsynergy uses expression data from ovarian, breast, and thyroid cancers sourced from The Cancer Genome Atlas (TCGA); forms MiRMs based on co-occurrence of miRNA targets; expands modules with a greedy include/exclude algorithm that maximizes a synergy score derived from miRNA–mRNA and gene–gene interactions; and applies a deterministic overlapping clustering algorithm adapted from a recently developed framework.