Introduction

SeqFold.jl is a high-performance pure Julia reimplementation of seqfold Python library for predicting nucleic acid secondary structures and calculating melting temperatures, which is, in turn, an implementation of the Zuker, 1981 dynamic programming algorithm, the basis for UNAFold/mfold, with energy functions from SantaLucia, 2004 (DNA) and Turner, 2009 (RNA).

Motivation

Secondary structure prediction is essential for:

• Designing PCR primers with minimal secondary structure;
• Creating oligos for genome editing techniques like MAGE;
• Tuning ribosome binding site (RBS) expression rates;
• Analyzing potential off-target binding in CRISPR applications.

SeqFold.jl provides a minimalist open-source alternative to proprietary solutions with on-demand access. The package has well-documented API and no dependencies other than standard Julia library.

For $T_m$ calculation the package provides fine-grained ionic conditions control.

Julia programming language is well-known for its speed and this package outperforms both the original Python package and Biopython Tm_NN reference function:

Basic Usage

Installation

First option (inside Julia REPL):

julia> ]
pkg> add SeqFold

Second option (from Julia scripts):

using Pkg
Pkg.add("SeqFold")

either way you should now be able to call the package:

using SeqFold

Melting Temperature Calculation

julia> seq = "GGGAGGTCAGCAAACCTGAACCTGTTGAGATGTTGACGTCAGGAAACCCT";

julia> tm(seq) # melting at PCR conditions, same as tm(seq, conditions=:pcr)
80.4

julia> MeltingConditions(:pcr) # here they are as a preset inherited from seqfold library
MeltingConditions (PCR preset)
  • NEB PCR buffer conditions for Taq DNA Polymerase
  • seq1 concentration: 250.0 nM (typical primer concentration)
  • seq2 concentration: 0.0 nM (asymmetric PCR)
  • Mg²⁺: 1.5 mM (optimal for Taq DNA Polymerase per NEB guidelines)
  • K⁺: 50.0 mM
  • Tris: 2.0 mM
  • dNTPs: 0.2 mM

julia> tm(seq, Mg=5) # altering default conditions
81.4

julia> MeltingConditions(:std) # second preset, inherited from seqfold library
MeltingConditions (standard preset)
  • Standard hybridization buffer
  • seq1 concentration: 25.0 nM
  • seq2 concentration: 25.0 nM
  • Na⁺: 50.0 mM

julia> tm(seq, conditions=:std, Na=150) # altering chosen preset
79.8

julia> custom_conds = MeltingConditions( # Specifying custom conditions
       seq1_conc=30,seq2_conc=20,Na=5,K=5,Tris=15,Mg=0,dNTPs=0)
MeltingConditions (custom)
  • seq1 concentration: 30.0 nM
  • seq2 concentration: 20.0 nM
  • Na⁺: 5.0 mM
  • K⁺: 5.0 mM
  • Tris: 15.0 mM

julia> tm(seq, conditions=custom_conds, Tris=50) # flexible adjustments
68.4

Secondary Structure Prediction

# Get the list of possible secondary structures for a sequence
julia> folded = fold(seq) # calculated at 37°C by default, Julia uses 1-based indexing
12-element Vector{SeqFold.Structure}:
    3   40   -1.3  STACK:GA/CT    
    4   39   -1.4  STACK:AGG/TGC  
    6   37   -1.5  STACK:GT/CA    
    7   36   -1.3  STACK:TC/AG    
    8   35   -1.5  STACK:CA/GT    
    9   34   -2.0  STACK:AGC/TTG  
   11   32   -1.5  STACK:CA/GT    
   12   31    2.8  INTERIOR_LOOP:4/2
   16   29   -1.3  STACK:CT/GA    
   17   28   -0.1  STACK:TGA/AGT  
   19   26   -1.0  STACK:AA/TT    
   20   25    3.5  HAIRPIN:AC/TG 

julia> dot_bracket(seq, folded) # get the dot-bracket notation
"..((.((((.((...((.((....)).)).)).)))).)).........."

julia> folded_hot = fold(seq, temp=80)
4-element Vector{SeqFold.Structure}:
    7   19   -0.4  STACK:TC/AG    
    8   18   -0.5  STACK:CA/GT    
    9   17   -0.4  STACK:AG/TC    
   10   16    3.6  HAIRPIN:GC/CC 

julia> dot_bracket(seq, folded_hot) # at a higher temperature number of secondary structures is decreased
"......((((.....))))..............................."

julia> dg(folded) # get the total free energy of the predicted secondary structures
-6.6

julia> dg(folded_hot)
2.3

julia> dg(seq, temp=4) # this method calls the computationally intensive `fold` internally
-16.2

Calling from Python scripts

Install juliacall package through you favourite package manager (pip, Conda, uv):

$ pip install juliacall

In Python REPL:

Python 3.13.7 | packaged by conda-forge | (main, Sep  3 2025, 14:30:35) [GCC 14.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> from juliacall import Main as jl # this will download Julia binaries
>>> jl.seval("""using Pkg; Pkg.add("SeqFold")""")
>>> jl.seval("using SeqFold") # this exports tm, fold, dg, ...
>>> jl.fold("ACACGCAGCGGGTGTGGCTGTGGTGC")
Julia:
4-element Vector{SeqFold.Structure}:
    1   13   -1.9  STACK_DE:AC/TG 
    2   12   -2.3  STACK:CAC/GGG  
    4   10   -2.2  STACK:CG/GC    
    5    9    3.5  HAIRPIN:GC/CG  
>>> jl.dg("ACACGCAGCGGGTGTGGCTGTGGTGC")
-2.9

Remember, that calling a function in Julia for the first time takes more time due to JIT compiliation.

Citations

Papers, which helped in developing the original seqfold package:

• Nussinov, 1980:

  Nussinov, Ruth, and Ann B. Jacobson. "Fast algorithm for predicting the secondary structure of single-stranded RNA." Proceedings of the National Academy of Sciences 77.11 (1980): 6309-6313.

• Zuker, 1981:

  Zuker, Michael, and Patrick Stiegler. "Optimal computer folding of large RNA sequences using thermodynamics and auxiliary information." Nucleic acids research 9.1 (1981): 133-148.

• Jaeger, 1989:

  Jaeger, John A., Douglas H. Turner, and Michael Zuker. "Improved predictions of secondary structures for RNA." Proceedings of the National Academy of Sciences 86.20 (1989): 7706-7710.

• SantaLucia, 2004:

  SantaLucia Jr, John, and Donald Hicks. "The thermodynamics of DNA structural motifs." Annu. Rev. Biophys. Biomol. Struct. 33 (2004): 415-440.

• Turner, 2009:

  Turner, Douglas H., and David H. Mathews. "NNDB: the nearest neighbor parameter database for predicting stability of nucleic acid secondary structure." Nucleic acids research 38.suppl_1 (2009): D280-D282.

• Ward, 2017:

  Ward, M., Datta, A., Wise, M., & Mathews, D. H. (2017). Advanced multi-loop algorithms for RNA secondary structure prediction reveal that the simplest model is best. Nucleic acids research, 45(14), 8541-8550.