circGPA

circGPA calculates associations between circular RNAs (circRNAs) and functional annotation terms by constructing miRNA-based interaction networks and using probability-generating functions to compute deterministic p-values for association tests.

Key Features:

• Deterministic Approach: Uses a deterministic algorithm instead of Monte-Carlo sampling to provide precise quantification of associations between circRNAs and annotation terms.
• Probability-Generating Functions: Employs probability-generating functions to calculate p-values for association tests between circRNAs and annotation terms.
• miRNA-based Interaction Networks: Builds interaction networks from known circRNA–miRNA interactions and leverages miRNA–mRNA interactions for annotation propagation.
• Computational Efficiency: Delivers computational performance up to two orders of magnitude faster than Monte-Carlo sampling methods, enabling rapid processing of large circRNA datasets.

Scientific Applications:

• Functional Annotation of circRNAs: Predicts functional annotations for circRNAs based on known interactions with miRNAs and downstream miRNA–mRNA interactions.
• Biomarker Discovery: Facilitates identification of disease-associated circRNA biomarkers by providing functional association annotations.
• Large-Scale Data Analysis: Supports summarizing and reannotating extensive circRNA databases through efficient association testing.

Methodology:

For each target circRNA, an interaction network is constructed from known circRNA–miRNA interactions; functional information from annotated nodes is propagated to the root circRNA node; association p-values between the circRNA and annotation terms are computed deterministically using probability-generating functions.