Multiomic network analysis reveals conserved and subtype-specific cooperative microRNA regulators in breast cancer

MicroRNAs (miRNAs) regulate gene expression through many-to-many interactions that are cooperative, context-dependent, and challenging to interpret from individual miRNA-target pairs alone. Yet, no framework simultaneously captures the cooperative, systems-level, and context-specific dimensions of miRNA regulation across the heterogeneous subtypes of a breast cancer. Here, we integrated paired miRNA and mRNA expression profiles from 1,161 breast cancers from The Cancer Genome Atlas (TCGA) with experimentally validated miRNA-target interactions and protein interaction networks to define cooperative miRNA regulatory modules in breast cancer. Community-resolved analysis of the resulting networks identified 16 functionally coherent miRNA modules associated with core cancer processes. To capture pathway-level network context beyond direct targets, we quantified the proximity of miRNA target sets to pathway proteins within a breast cancer-specific protein interaction network. This framework identified cooperative miRNA modules associated with cell cycle progression, DNA repair, PTEN/TP53-related signaling, and epithelial-mesenchymal transition. Subtype-resolved network analysis further showed that most inferred miRNA functions are strongly context dependent, not being associated in all of basal-like, HER2-enriched, luminal A, and luminal B tumors. In contrast, a limited set of pathway-level associations was conserved across subtypes. Among these, the miR-29 family (miR-29a/b/c-3p) emerged as a consistent regulator of collagen remodeling, extracellular matrix organization, PDGF signaling, and EMT. Notably, this functional conservation persisted despite subtype-specific variation in the underlying target genes. Together, these results define the cooperative miRNA modules that post-transcriptional regulate breast cancer subtypes, identify the miR-29 family as a candidate pan-subtype therapeutic target for EMT-driven metastasis, and establish a generalizable computational framework for resolving miRNA programs across heterogeneous cancers.