IPknot

IPknot predicts RNA secondary structures including pseudoknots to model complex RNA folding for functional and comparative analyses.

Key Features:

• Maximizing Expected Accuracy (MEA): Implements an MEA criterion to select secondary-structure predictions that maximize expected base-pairing accuracy.
• Decomposition into pseudoknot-free substructures: Decomposes complex pseudoknotted structures into simpler pseudoknot-free substructures for tractable modeling.
• Approximate base-pairing probability distribution: Approximates a base-pairing probability distribution that accounts for pseudoknots.
• Integer Programming with threshold cut: Solves the MEA optimization using integer programming (IP) with a threshold cut strategy.
• Heuristic algorithm for refinement: Applies a heuristic algorithm to refine base-pairing probabilities and improve prediction precision.
• Consensus secondary structure prediction: Predicts consensus secondary structures with pseudoknots from multiple sequence alignments.

Scientific Applications:

• Functional RNA analysis: Facilitates analysis of functional RNAs by predicting secondary structures that include pseudoknots to inform functional interpretation.
• Comparative genomics: Enables comparative analyses and evolutionary studies via consensus secondary-structure prediction from multiple sequence alignments.

Methodology:

IPknot approximates base-pairing probability distributions that include pseudoknots, decomposes pseudoknotted structures into pseudoknot-free substructures, maximizes expected accuracy (MEA), solves the MEA optimization via integer programming with threshold cut, applies a heuristic algorithm to refine base-pairing probabilities, and supports consensus prediction from multiple sequence alignments.