miRNALasso

miRNALasso applies adaptive Lasso regularized regression to integrate sequence features and miRNA and mRNA expression profiles to predict functional miRNA-mRNA interactions and quantify miRNA down-regulation effects.

Key Features:

• Sequence features: Uses context features of target sites, thermodynamic stability, and accessibility energy to characterize miRNA-mRNA interactions.
• Expression profile analysis: Integrates miRNA and mRNA expression profiles to assess regulatory effects alongside sequence information.
• Adaptive Lasso regularized regression: Employs adaptive Lasso procedures to quantify the down-regulation effect of each miRNA and to estimate contributions of individual sequence features.
• Improved sensitivity and specificity: Demonstrated superior sensitivity and specificity on expression datasets from cancer studies compared to other existing methods.
• Comprehensive feature integration: Integrates diverse sequence features rather than treating putative miRNA-mRNA interactions as equally significant.

Scientific Applications:

• Cancer miRNA target discovery: Identification and quantification of functional miRNA targets in cancer expression datasets.
• Regulatory network analysis: Elucidation of miRNA-mediated regulatory networks and the roles of miRNAs in biological processes and disease pathogenesis by combining sequence and expression data.

Methodology:

Constructs a predictive model that simultaneously considers gene sequence features and miRNA/mRNA expression profiles using adaptive Lasso regularized regression to predict targets and quantify regulatory impact.