-p (predictor) parameter selects which method to use in step 2 of the FoldARE algorithm.

Available options:

• E - EternaFold (machine learning-based)
• R - RNAstructure (thermodynamics-based) - default
• L - LinearFold (linear-time algorithm)
• V - ViennaRNA (thermodynamics-based)

The predictor takes the pseudo-SHAPE co-input generated from step 1 to guide the final structure prediction.

Default: R (RNAstructure) - performed best in our benchmarks when combined with EternaFold as ensembler.

-e (ensembler) parameter selects which method to use in step 1 of the FoldARE algorithm.

Available options:

• E - EternaFold (machine learning-based) - default
• R - RNAstructure (thermodynamics-based)
• L - LinearFold (linear-time algorithm)
• V - ViennaRNA (thermodynamics-based)

The ensembler generates a structural ensemble (set of alternative conformations), which is parsed to calculate per-position single-strand frequency f(ssRNA).

Default: E (EternaFold) - performed best in our benchmarks when generating ensembles for pseudo-SHAPE derivation.

--ens_n (ensemble size) controls the number of alternative structures generated in step 1.

The ensembler generates this many structures, ordered from most to least energetically favored (Zuker ordering).

Effect on prediction:

• Larger ensembles provide better statistics for f(ssRNA) calculation
• Smaller ensembles run faster but may miss rare conformations
• Default (75) was optimized for best performance/speed trade-off

Requirements: Integer greater than 1

Default: 75 (optimized value from our parameter search)

--shape allows you to provide custom SHAPE reactivity data to guide structure prediction.

What it does:

• When enabled, uses your provided SHAPE data instead of automatically generating pseudo-SHAPE from the ensemble
• When disabled (default), FoldARE automatically generates pseudo-SHAPE data from the ensemble

SHAPE data format:

• Tab-separated values
• Column 1: nucleotide position (1-indexed)
• Column 2: reactivity value (typically 0-2, normalized)
• Column 3: optional notes/comments

Example:
1 0.125
2 0.345
3 0.678
4 1.234
5 0.456

Note: Reactivity values should correspond to your input sequence length.