RNAup

What is RNAup?

RNAup calculates the thermodynamics of RNA-RNA interactions by accounting for the cost of making binding sites accessible. Unlike simpler duplex prediction methods, RNAup recognizes that potential binding regions in both RNA molecules are often base-paired within their own secondary structures. Before two RNAs can interact, these intramolecular structures must be disrupted, which costs energy.

The total binding energy RNAup reports combines two components:

$\Delta G_{\text{total}} = \Delta G_{\text{interaction}} + \Delta G_{\text{opening}}$

The interaction energy (ΔG_interaction) represents the stabilizing base pairing between the two molecules. The opening energy (ΔG_opening) represents the energetic cost of unpairing the binding sites in one or both molecules so they become available for intermolecular pairing. This accessibility-aware approach produces more biophysically realistic predictions than methods that ignore the structural context of binding sites.

Applications

Accessibility controls the efficacy of many RNA-mediated processes. MicroRNAs, siRNAs, and bacterial small RNAs all require their target sites to be accessible for efficient binding. If seed sequences are buried within stable secondary structure, the RNA-induced silencing complex (RISC) cannot access them effectively. Studies have shown that target site accessibility correlates directly with siRNA knockdown efficiency and miRNA repression strength.

RNAup predictions are useful for:

• siRNA and miRNA target validation: Predicting whether a target site is accessible enough for efficient silencing
• sRNA target identification in bacteria: Finding mRNA targets of regulatory small RNAs
• Antisense oligonucleotide design: Selecting accessible binding sites for therapeutic applications
• Ribosome binding site analysis: Evaluating whether Shine-Dalgarno sequences are accessible for translation initiation

How does RNAup work?

RNAup uses partition function calculations over all possible conformations to determine the probability that a binding site remains unpaired in equilibrium. This probability p_u relates to opening energy via:

$\Delta G_{\text{opening}} = -RT \ln(p_u)$

The algorithm operates in two stages. First, it computes accessibilities for regions of specified length, reporting the free energy required to open each potential binding site. Second, it combines these opening energies with duplex interaction energies to find the optimal interaction.

An efficient O(n³) algorithm computes accessibilities for all intervals within a sequence simultaneously. Earlier implementations could only compute intervals of one fixed length in the same time. This improvement enables scanning for binding sites across a range of interaction lengths without repeated computation.

How to use RNAup online

ProteinIQ provides browser-based access to RNAup, eliminating the need to install the ViennaRNA package locally.

Inputs

Settings

Output

Results appear in a spreadsheet with columns for interaction details.

Interpreting results

More negative total energy values indicate stronger predicted interactions. When comparing potential binding sites, consider both the total energy and the relative contributions of interaction and opening energy. A site with favorable duplex energy but high opening cost may be less effective than a site with slightly weaker base pairing but greater accessibility.

Limitations

RNAup assumes that both RNA molecules fold independently before interaction. It does not model cooperative refolding where the interaction itself might stabilize alternative structures. For very long sequences, computation time increases significantly due to the O(n³) complexity.

The method predicts thermodynamic favorability, not kinetic accessibility. A binding site might be thermodynamically accessible but kinetically trapped behind folding barriers. Experimental validation remains essential for confirming predicted interactions.

Related tools

• RNAduplex: Simpler duplex prediction without accessibility. Faster but less accurate for structured RNAs.
• RNAplex: Fast interaction prediction with accessibility. Better for screening many potential interactions.
• RNAcofold: Predicts the full secondary structure of an RNA-RNA dimer complex.
• RNAplfold: Computes local base pair probabilities and accessibility profiles for long sequences.
• RNAfold: Single-sequence structure prediction. Use to examine the secondary structure of individual RNAs before interaction analysis.
• ViennaRNA: Access all 14 ViennaRNA methods through a unified interface