RNAsubopt

What is RNAsubopt?

RNAsubopt generates all RNA secondary structures within a specified energy range above the minimum free energy (MFE). Unlike RNAfold, which returns only the single most stable structure, RNAsubopt exhaustively enumerates alternative conformations that are thermodynamically plausible.

This exhaustive approach matters because RNA molecules do not exist in a single conformation. At physiological temperatures, an RNA population samples many structures from its Boltzmann ensemble, and functionally important conformations may not be the MFE structure. RNAsubopt reveals this structural diversity.

The algorithm was first described by Wuchty et al. (1999) and implemented in the ViennaRNA Package. It uses dynamic programming to systematically trace back through the energy landscape, collecting all structures whose free energy falls within the specified delta of the optimal.

When to use RNAsubopt

RNAsubopt is particularly useful for:

• Riboswitch analysis: Riboswitches function by adopting alternative structures in response to ligand binding. Enumerating suboptimal structures reveals these functional conformations.
• Kinetic accessibility studies: Structures close in energy to the MFE may be kinetically more accessible than the global minimum.
• Functional site accessibility: Determining whether regulatory sites like Shine-Dalgarno sequences are exposed in the ensemble.
• Folding kinetics: Combined with the barriers program, suboptimal structures enable analysis of RNA folding pathways.

How to use RNAsubopt online

Paste one or more RNA sequences, set an energy range above the MFE, and RNAsubopt returns every secondary structure within that window. Results appear as a sortable table of dot-bracket structures with their free energies, ready for download.

Input

Field	Description
RNA Sequences	One or more RNA sequences in FASTA format or as plain text. Accepts .fasta, .fa, or .txt files up to 10 MB.

Pricing and runtime

RNAsubopt uses runtime billing at 36 credits per minute, rounded up to a whole credit, with an 8-credit minimum for successful jobs. The spending limit is shown before submission. Credits are reserved in stages, and unused reserved credits are returned after settlement. The same pricing applies when RNAsubopt is selected in ViennaRNA.

Jobs can use up to two hours of compute, subject to the spending limit and available credits. A 30-minute run uses 1,080 credits; two hours uses 4,320 credits. Reaching the spending limit stops the calculation and charges the reserved runtime credits. Execution failures retain the existing refund policy.

Exhaustive enumeration has an unpredictable runtime. The initial estimate does not guarantee completion within that time or budget. Large output can also exhaust memory or exceed the result publication limit before it can be downloaded.

Settings

Suboptimal enumeration

Setting	Description
Energy range	Optional delta above MFE in kcal/mol. A blank value uses ViennaRNA 2.7.2's native 1 kcal/mol range. Larger values can yield exponentially more structures.
Maximum structures	Optional number of rows returned per sequence. A blank value returns all rows. This is applied after native enumeration, so it does not shorten the calculation or cap the native output file.

Prediction options

Setting	Description
Temperature	Folding temperature in Celsius (0-100, default 37). Affects energy calculations and which structures fall within the range.
Disallow lonely pairs	When enabled, excludes structures containing isolated base pairs (helices of length 1). Reduces output volume and may improve biological relevance.
Dangling ends	Treatment of unpaired nucleotides adjacent to helices. Double dangles (default) is recommended for most applications.

Output

Results are returned as a table with one row per enumerated structure:

Column	Description
Sequence ID	Identifier from FASTA header or auto-generated.
Structure #	Position in native output order. Structures are not sorted by energy.
Length	Sequence length in nucleotides.
Secondary Structure	Dot-bracket notation showing base pairs.
Energy	Free energy in kcal/mol.
Is MFE	Whether this row has the lowest energy among the returned rows. A row cap can exclude the native MFE structure.

Limitations

The number of suboptimal structures grows exponentially with both sequence length and energy range. For a 50-nucleotide sequence, an energy range of 5 kcal/mol might yield hundreds of structures, while the same range on a 100-nucleotide sequence could produce millions.

Practical constraints:

• Long sequences and wider energy ranges can produce more structures than can be enumerated within the runtime and spending limits
• The Maximum structures setting reduces returned table rows after computation; it does not prevent memory growth during enumeration
• Very large energy ranges (>10 kcal/mol) are rarely useful and dramatically increase computation time

For longer sequences, consider RNAplfold for local structure analysis, or use RNAsubopt with the Disallow lonely pairs option enabled to reduce output volume.