What is RNAinverse?
RNAinverse solves the inverse RNA folding problem: given a target secondary structure in dot-bracket notation, it designs RNA sequences that fold into that structure. Introduced in the ViennaRNA package over 20 years ago, RNAinverse was the first computational approach to RNA sequence design and remains one of the fastest methods available.
The algorithm uses an adaptive random walk that iteratively mutates a candidate sequence, evaluating each mutation by comparing the predicted minimum free energy (MFE) structure against the target. By default, the search terminates when a sequence folds exactly into the desired structure.
Applications
RNA inverse folding enables practical applications across synthetic biology and research:
- Synthetic RNA design: Engineering riboswitches, aptamers, and other functional RNAs with specific structural requirements
- mRNA optimization: Designing 5' UTR structures to control translation efficiency
- Biosensor development: Creating RNA molecules that adopt target conformations in response to ligands
- Experimental validation: Generating diverse sequence variants that fold identically for mutagenesis studies
How to use RNAinverse online
ProteinIQ provides cloud-based access to RNAinverse, eliminating the need for local installation of the ViennaRNA package.
Inputs
| Input | Description |
|---|---|
Target Structure | Secondary structure in dot-bracket notation. Parentheses denote base pairs: ( pairs with the corresponding ). Dots represent unpaired nucleotides. |
Structure notation example: ((((....)))) represents a stem-loop with 4 base pairs and a 4-nucleotide loop.
Settings
| Setting | Description |
|---|---|
Number of designs | How many independent sequence designs to generate (1-20, default 5). Each design starts from a different random sequence. |
Temperature | Folding temperature in Celsius (0-100, default 37). Affects thermodynamic calculations. |
Results
RNAinverse returns a spreadsheet with one row per designed sequence:
| Column | Description |
|---|---|
Design # | Sequential number identifying each design attempt |
Designed Sequence | The RNA sequence generated by the algorithm |
Actual Structure | The MFE structure that the designed sequence folds into |
MFE (kcal/mol) | Minimum free energy of the folded structure |
Distance | Base pair distance between the actual structure and the target (0 = perfect match) |
Matches Target | Whether the designed sequence folds exactly into the target structure |
How RNAinverse works
RNAinverse performs an adaptive random walk through sequence space. Starting from a random nucleotide sequence, the algorithm:
- Predicts the MFE structure of the current sequence using the same thermodynamic model as RNAfold
- Calculates the base pair distance between the predicted structure and the target
- Mutates nucleotides that are involved in incorrect base pairs
- Accepts mutations that reduce the distance to the target
- Repeats until the distance reaches zero or the search stalls
The base pair distance counts the number of base pairs that differ between two structures. A distance of zero means the designed sequence folds exactly into the target.
Interpreting results
Not all design attempts succeed. A Distance value greater than zero indicates the algorithm could not find a sequence folding exactly into the target. This occurs more frequently with:
- Complex structures: Large structures with many stems and junctions are harder to design
- Thermodynamically unfavorable targets: Some structures cannot be the MFE fold for any sequence
- Pseudoknots: RNAinverse only handles nested structures; pseudoknotted targets are not supported
Multiple designs with distance zero provide sequence diversity while maintaining the target fold. These sequences can be useful for experimental studies requiring structural equivalence with sequence variation.
Limitations
RNAinverse does not guarantee a solution exists for every target structure. Some structures are thermodynamically impossible to achieve as the MFE fold. If all design attempts return non-zero distances, consider whether the target structure is realistic.
The algorithm optimizes for the MFE structure only. A designed sequence might fold into the target as its MFE, but have many near-optimal alternative structures. For applications requiring a well-defined structural ensemble, verify the design using RNAsubopt to examine alternative conformations.
Related tools
- RNAfold: Predict secondary structure from sequence (the forward problem that RNAinverse inverts)
- RNAeval: Calculate the free energy of a sequence-structure pair
- RNAsubopt: Enumerate suboptimal structures to assess structural specificity of designed sequences
- RNAplot: Visualize secondary structures in graphical format
- RNAdos: Iterative inverse folding with more thorough optimization
- ViennaRNA: Access all 14 ViennaRNA methods through a unified interface